******************************************************************************* Content Manager V7.1.0 Fixpack 20 for AIX January 26,2003 This is an interim fixpack between Formal Corrective Service Deliverables (CSDs). Table of Contents 1. Installation Requirements 2. Installation Instructions 3. APAR Descriptions 4. Documentation APAR Fixes with Descriptions ******************************************************************************* ********************************************* 1. Installation Requirements ********************************************* In order to install this fixpack, you must have Content Manager v71 installed (syslevel 7.1.0.0) To determine your AIX syslevel, enter: lslpp -L frn.* ********************************************* 2. Installation Instructions ********************************************* 1. Download the cm710.20.tar.gz file in binary mode to a temporary subdirectory on your workstation. You will need at least 20mb of free space. 2. To unpack the GZIP compression, enter: gzip -d cm710.20.tar.gz If you don't have GZIP, you can obtain this freeware tool from: http://aixpdslib.seas.ucla.edu/aixpdslib.html 3. Unpack the cm710.20.tar file using the AIX tar utility. For example: tar -xvf cm710.20.tar 4. Log in as root. 5. Start the Installation wizard by entering ./frnxupdate.sh 6. After the installation has completed, if you are using DB2 must rebind the databases of your library server and object server. You can rebind the server databases as follows: Library Server: logon as your library server admin id (lsadmin by default) Issue the command: frnbind.lib Object Server: logon as your object server admin id (osadmin by default) Issue the command: frnbind.obj 7. CM 7.1 AND ORACLE SUPPORT In addition to existing product support for Oracle 8.0.3 and Oracle 8.1.6, fixpack 12 also provides support for Oracle 8.1.7.4. Please note that support for Oracle 8.1.7 is limited to Oracle 8.1.7.4 as a result of Oracle's urgence to all their customers to move to Oracle 8.1.7.4. Oracle 8.1.7.4 resolves known stability and reliability issues with previous versions of Oracle 8.1.7. Support for previous versions of Oracle 8.1.7 has been deprecated. Although we are not presently aware of any specific issues with other versions of Oracle, customers should note that testing has specifically been done only with Oracle 8.0.3, 8.1.6, and 8.1.7.4. Please also note that support refers to the machine on which the Library Server and Object Server codebase are installed. Thus, if the Library Server or Object Server database is not on the same machine as the Library Server or Object Server codebase, customers should assume these instructions apply to the machine where the codebase is actually installed. Accordingly, the level of Oracle which exists on the same machine as the Library Server or Object Server codebase must meet CM's minimum requirements for the level of Oracle you wish to support. For example, if you are running the Library Server codebase against a remote Oracle Library Server database and wish to enable CM's Oracle 8.1.7 support, the Oracle client machine must be at Oracle 8.1.7.4 client level and have Oracle Programmer components installed. The level of the Oracle Server is not relevant for CM purposes. Similarly, if you are running the Library Server codebase on a machine that also has the Library Server database (that is, Library Server codebase running on Oracle Server machine) and wish to enable CM's Oracle 8.1.7 support, you must be running at Oracle 8.1.7.4 server level with Oracle Programmer components installed. As always, you should be aware of any potential Oracle conflicts with Oracle client/server levels as well as any other known Oracle issues which may affect your environment. It is critical that Oracle patchsets be applied completely and correctly, otherwise unpredictable results may occur. For existing databases, there are often manual steps required by Oracle that must be taken after the patchset is installed to bring existing databases up to the level of the patchset. This is known to be the case for the Oracle 8.1.7.4 patchset and may be true for other patchsets as well. Follow Oracle's instructions completely and correctly to ensure your CM databases and Oracle product are at the proper Oracle level before following the procedures below. AIX ORACLE SUPPORT The default libraries installed during an Oracle installation of CM are compatible with Oracle 8.0.3 as well as Oracle 8.1.6. Thus, unlike Windows platforms, there are no additional steps required to enable Oracle 8.1.6 support on AIX platforms. We recommend you make sure there are no clients currently connected to CM before following these procedures. Oracle 8.1.7.4 support 1. Stop the CM Library Server, Object Server, and SMS Server. To do this, under the appropriate CM administrator userid, run "frnxcmd shutdown" followed by "frnkill.all". Stopping the Object Server will also stop the SMS Server. 2. Run "ps -ef | grep -i frn" to ensure there are no server processes still active. 3. Login as root, perform an slibclean to ensure all libraries are removed from memory. 4. Locate and change to the directory where you untar'd the CM fixpack. 5. If you are enabling Oracle 8.1.7.4 support, copy all the *.o files in ./Orautil/817 to your /usr/lpp/frn/lib directory after making a backup of your originals. The ./Orautil/817 directory contains the following files: frnxlblb.o Object Server Shared Library frnxllff.o Library Server Shared Library frnxllio.o Library Server Shared Library frnxllls.o Library Server Shared Library frnxlluq.o Library Server Shared Library frnxmiut.ORACLE Library Server Migration Utility 6. If the Object Server and Library Server are on separate AIX machines, the respective DLL's should be copied to each machine. 7. Once the Shared Libraries have been copied, you may re-start the CM services. Check frndiag.log to ensure they have started successfully. 8. The frnxmiut.ORACLE file is not required for normal CM operation. To be used only if you are migrating the Library Server from an earlier version of CM 7.1. Oracle 9.2.0.4 support 1. Stop the CM Library Server, Object Server, and SMS Server. To do this, under the appropriate CM administrator userid, run the "frnstop" utility. Stopping the Object Server will also stop the SMS Server. 2. Run "ps -ef | grep -i frn" to ensure there are no server processes still active. 3. Login as root, perform an slibclean to ensure all libraries are removed from memory. 4. Locate and change to the directory where you untar'd the CM fixpack. 5. If you are enabling Oracle 9.2.0.4 support a. Backup the *.o files in your /usr/lpp/frn/lib directory b. Copy all the *.o files in ./Orautil/920 to your /usr/lpp/frn/lib directory. The ./Orautil/920 directory contains the following files: frnxlblb.o Object Server Shared Library frnxllff.o Library Server Shared Library frnxllio.o Library Server Shared Library frnxllls.o Library Server Shared Library frnxlluq.o Library Server Shared Library frnxmiut.ORACLE Library Server Migration Utility 6. Because Oracle 9i is a 64-bit application, you should recompile your index class access modules to ensure they will be both 9i and 32-bit compatible. Login with your Library Server administrator user id and perform the following steps: export ORACLE_SID= sqlplus / UPDATE SBTCLASSDEFS SET DLLSTATUS = -1 WHERE DLLSTATUS = 0; UPDATE SBTVIEWDEFS SET DLLSTATUS = -1 WHERE DLLSTATUS = 0; COMMIT; exit 7. If the Object Server and Library Server are on separate AIX machines, the respective DLL's should be copied to each machine. 8. Once the Shared Libraries have been copied, you may re-start the CM services. Check frndiag.log to ensure they have started successfully. 9. The frnxmiut.ORACLE file is not required for normal CM operation. To be used only if you are migrating the Library Server from an earlier version of CM 7.1. 8. LIBRARY SERVER MIGRATION UTILITY ENHANCEMENTS FOR CM FIXPACK 13 This fixpack introduces two major enhancements to improve the Library Server migration experience for customers an extensive, context-sensitive logging file that details all actions taken by the migration utility, and ability to create missing system index classes for those who initially migrated from VI/DL 2.4 or earlier on OS/2 systems. As before, this Library Server migration utility supports migrations from VI/DL 2.4 for Windows or AIX. For OS/2 systems, your Library Server database must be at VI/DL 2.4 level or higher to use this migration utility. Note that there are a total of 7 versions of this utility in the fixpack. DBMS OPERATING SYSTEM FIXPACK FULL PATH DB2 7.1 or higher Windows \migrate\frnnmiut.DB2.exe DB2 7.1 or higher AIX \migrate\frnxmiut.DB2 Oracle 8.0.3 Windows \migrate\frnnmiut.ORACLE.exe Oracle 8.1.6 Windows \Orautil\816\frnnmiut.ORACLE.exe Oracle 8.1.7.4 Windows: \Orautil\817\frnnmiut.ORACLE.exe Oracle 8.0.3 AIX \migrate\frnxmiut.ORACLE Oracle 8.1.7.4 AIX \Orautil\817\frnxmiut.ORACLE Please note that, as before, the migration utility should be copied into a local directory. For DB2 installations, this must also include the frnomiut.bnd bind file and the two files must reside in the same directory. The userid who will run the utility must have read/write permission for the local directory in which the migration utility was copied. Further, the userid must have the necessary privileges to logon to the Library Server database and perform various operations such as database create, update, insert, and delete operations. We recommend you use the Library Server administrator account. For Oracle systems, your tnsnames and listener files must be properly configured. To verify this, you should be able to logon to the Library Server database using sqlplus with the Library Server administrator userid and password. Assuming your ORACLE_SID is set for the Library Server database, you should be able to logon by typing "sqlplus /" from the command prompt. If you encounter errors while verifying this, consult your Oracle DBA or Oracle documentation for further assistance. Logging enhancements: Debugging a failed migration attempt has been made easier through the addition of extensive logging. A file, frnpmiut.log, will be written in the same local directory to which you copied the migration utility executable. This file contains a complete record of actions performed by the migration utility. It includes the action performed, SQL error returned, contextual result message (alerting you to when an error is expected or can be safely ignored), the actual SQL statement being executed, and the actual DBMS error message associated with the action (not available for Oracle AIX systems). This file will be automatically created and should be analyzed when experiencing migration difficulties. Oracle-specific Notes: For Oracle systems, please note also that there are some situations where a failed migration attempt cannot be recovered. As previously recommended, it is important that you retain a backup of your existing CM Library Server database until you are confident that the migration has completed successfully. The logfile will alert you to situations in which it could not recover from a failed migration. Users who plan to run this utility as the result of a previously failed migration attempt are strongly encouraged to restore their pre-migration Library Server database. System Index Class Toleration enhancements: Typing the migration utility filename with no command line inputs, will result in a display similiar to the following: Usage: frnnmiut.DB2 [toleration] (e.g.: frnnmiut.DB2 libsrvrn 0) = (Required) Library Server database name. = (Required) Oracle schema owner/userid for LS database. = (Required) Oracle password for userid above. [toleration]= (Optional) For systems originally migrated from OS/2. Toleration settings are as follows (0 = default): 0 = abort on missing system-level index classes (no toleration) 1 = ignore missing system-level index classes (min toleration) 2 = create missing system-level index classes (max toleration) 3 = create missing system-level index classes *ONLY* NOTE: Toleration option #3 is ONLY for customers who have already successfully migrated to CM 7.1. If your Library Server database is not yet at a CM 7.1 level, you should only choose toleration options 0, 1, or 2. If you accidentally choose option 3 before migrating your Library Server database to CM 7.1, you will not be to choose any other toleration option and will need to restore your Library Server database backup. As the usage description implies, the migration utility is now able to tolerate CM systems whose databases were originally based on VI/DL 2.4 or earlier systems for OS/2. In some cases, customers may not have system index classes that are now installed by recent versions of CM systems. These system index classes allow you to perform such tasks as scheduling CM utilities, make use of Media and Search index classes, and a sample index class. Specifically, the system index classes which can be created by the migration utility are FRN$SEARCH, FRN$MEDIA, FRN$UTSCH, and DLSAMPLE. To determine if you need the additional index classes, logon to your Library Server database from a DB2 command prompt or Oracle SQL*Plus and issue the following query. If the results do not include the four index classes listed in the preceding paragraph, you should allow the migration utility to create them for you. SELECT KEYWORDSTRING FROM SBTNLSKEYWORDS WHERE KEYWORDCLASSFI=2 The toleration options are described as follows: 0 or no option specified - abort on missing system-level index classes (no toleration) The migration utility acts as it has in previous CM 7.1 levels. If any of the four system index classes mentioned above do not exist, an error will be thrown. the migration utility will log the error, output an error to the screen, and terminate. 1 - ignore missing system-level index classes (min toleration) This option provides the least amount of toleration by allowing a migration to proceed even if the four system index classes mentioned earlier do not exist. In this case, the absence of any such index classes will simply be ignored. A message, however, will be logged in frnpmiut.log. 2 - create missing system-level index classes (max toleration) This option provides the greatest flexibility and is the recommended option for all users who need to migrate from CM 6.1 or earlier. In addition to performing a normal CM migration, any of the four system index classes mentioned above which are found to be missing will be created and automatically migrated. The frnpmiut.log file will indicate that such a creation has taken place. 3 - create missing system-level index classes *ONLY* This option is ONLY for users who have already successfully migrated their Library Server to CM 7.1. This option allows you to utilize the index classes described above without the need to re-migrate your system. Note that, if you already have a working CM 7.1 system, it is recommended that you back up your CM 7.1 database before proceeding. In the event of a failure, you can restore your CM 7.1 database and choose this option again, without having to perform a full migration. If your Library Server is NOT at CM 7.1 level and you choose this option before migrating your system, you will not be able to choose the other migration options later. To do so, you will need to restore the backup of the Library Server database and choose one of the other options first. ********************************************* 3. APAR Fixes with Descriptions ********************************************* ************************************ Fixpack 710.20 ************************************ No APAR Fixes ************************************* ************************************ Fixpack 710.19 ************************************ APAR NUMBER: IR51892 APAR Description: Remote Object Server Migration fails to migrate objects. This only happens with objects that have previously been replicated. The problem is seen on the library server as the failure occurs when the migrator tries to update the SBTCOPIES table. Fix Description: Initialize the value of sqlca.sqlcode in functionLdbUpdateCopiesObjCodeAndLoc2. Special Instructions: none ************************************* APAR NUMBER: IR52383 APAR Description: Content Manager Object Server fix for Migrator problems Signal 11 on processing undo records when network problems can CM Server communications to drop. Fix Description: The migrator logic has been modified to generate correct undo record. Special Instructions: none ************************************* APAR NUMBER: IR52435 APAR Description: CM Object Server Purger DB query for Object is causing Oracle Undo logs to increase and could become an issue if the DB is not monitored and tuned. Fix Description: Modified the query to reduce the load on Oracle DB. Special Instructions: none ************************************* APAR NUMBER: PQ77783 APAR Description: For the CM for Z/OS 2.3 with the 7.1 modifications applied, the user is getting an 0C4 when using the EIP E-Client when the number of rows to be retrieved is more than 50. It can also happen if the columnlimit field is set to a large number like 255. This did not happen with the windows client. Fix Description: Change the type of the parameter in LLsAllocate for calculating the size for column description information from unsign short interger to unsigned long interger Special Instructions: none ************************************* APAR NUMBER: IR43518 APAR Description: Current documentation does not outline the requirements or procedures necessary to install multiple Digital Library systems from a remote site. Fix Description: Information on how to isntall and configure multiple systems is available on the IBM Support website. The link to the information is: http://www-1.ibm.com/support/search.wss?q=Version+7%2E1+installing+and+configuring+remotely&tc=SSRS7Z+&rs=86 The title of the document is: "Content Manager Version 7.1: Installing and configuring the servers remotely" Special Instructions: none ************************************* APAR NUMBER: IR50986 APAR Description: Online documentation for creating a database index on an exist- ing index class is incorrect. Symptom is that the system admin will return an FRN8727a (the value of the indexname parameter is not valid). Fix DESCRIPTION: The correct guidelines for the index name are: 1) Maximum is 18 characters. 2) Must start with a letter Can include uppercase letters, numbers, and underscores Cannot be a reserved word Special Instructions: none ************************************* APAR NUMBER: IR51254 APAR Description: Copying an index class in the left hand pane tree does not show the copy in the right hand pane: 1. Logon to the sysadmin client 2. natigate down the left hand tree to the index classes 3. choose to copy an existing index class 4 . rename the index class copy to something unique 5. click ok 6. the index class will not appear in the right hand window pane....refresh does not work. Fix Description: Set cache flag off in SajcNodeFRIndexClasses. This will allow the user to refresh the index class notes in the tree view. Special Instructions: none ************************************* APAR NUMBER: IR51375 APAR Description: The Cachecode returned in the LibGetReplyInfo structure is used to supply information to your application and should not be considered an error code. Fix Description: The CacheCode (returned by the LibGetReplyInfo structure) is used to supply information to your application, and should not be considered an error code. CacheCode represents the version number of the object. For example, a cache code of 1 means that the version number of the object/part is 1. Similarly, CacheCode = 2 means that the version number of the object/part is 2. Special Instructions: none ************************************* APAR NUMBER: IR51643 APAR Description: BUILDING FRNXLCPW LOGON USEREXIT ON AIX Fix Description: Before compiling the frnxlcpw logon RACF user exit for AIX, you must add this environment variable: export PROFILEOFF=1. Otherwise, you will receive the following errors when you compile: Compilation is failing with the following error message /usr/vacpp/bin/xlC_r -I /usr/lpp/frn/include -I . -O3 -qstrict -pg -qcpluscmt -qcpluscmt -c -DAIX -DTOOLKIT -DUSERAPP -qalign=packed -DBIT32 -DDLL -DSINGLETHREAD -DNOCACHE -o../obj/frnplcpw.o ./frnplcpw.c .... .... /usr/vacpp/bin/makeC++SharedLib_r -p$$ -lc -lbsd -lfrnxtk -L/usr/lpp/frn/lib -bmap:../obj/frnxlcpw.map ../obj/frnplcpw.o ld: 0711-317 ERROR: Undefined symbol: ._mcount ld: 0711-344 See the loadmap file ../obj/frnxlcpw.lmap for more information. make: 1254-004 The error code from the last command is 8. Special Instructions: none ************************************* APAR NUMBER: IR51747 APAR Description: Folder Manager API's and Threading. Fix Description: I have a created the a document on the e-support Web site, here's the URL, where you can verify the information and close this defect: http://www-1.ibm.com/support/docview.wss?uid=swg21137159 Special Instructions: Multiple change SMS orders on the same object are not supported. ************************************* ************************************ Fixpack 710.18 ************************************ APAR NUMBER: IR49901 APAR Description: CM Object Server command utility not handling lengthy GECOS information on user description_ Fix Description: Modified UNIX calls for user information to correct for buffer problems Special Instructions: none ************************************* APAR NUMBER: IR51552 APAR Description: The FRNCMD would freeze when used as a monitoring tool, meaning it was run every few minutes to verify the library server was up. Always failed in the same "validation" step. Fix Description: Only do the validation step for shutdown. Remove the validation check for STATUS Special Instructions: Be sure when compile this code the profiling option is off. ************************************* APAR NUMBER: IR51609 APAR Description: CM7.1 OBJECT SERVER REPLICATOR DOES NOT COMMIT OFTEN ENOUGH Fix Description: Default limit of 32 replications between replicator commits. Tuneable by setting environment variable FRNREPCOUNT to a integer value. Special Instructions: none ************************************* APAR NUMBER: IR51694 APAR Description: eClient application gets a Return Code 1 on a SimLibGetItemSnapshot API call. This is after applying the fix for IR51349. This customer is using eClient 8.1 and EIP 8.1 to talk with CM 7.1 servers, but any CM 7.1 client could experience this condition. Fix Description: Make the changes for Fixed_char type on pack, unpack area to get the correct row size and column size to pass to Client Special Instructions: To improve performance, indexes need to be defined on attribute column for all MVT table ************************************* APAR NUMBER: IR51700 APAR Description: If a system is configured to both replicate and migrate objects, there is a possibility of conflicts occurring between the migrator and the replicator. The migrator will not migrate an object if it is in the base_replication table. While reviewing a trace of an object server with this setup, it was noticed that this conflict sometimes results in an SQL return code of -502 in the migrator. A -502 is a "Cursor already open" error and should not occur. The result of this error just some additional overhead processing to correct the open cursor problem, there is no loss of data or other interference with migration or replication. Fix Description: Migrator is checking if the replication is in progress, if it is migrator will delay it operation, but need to roolback the cursor for DB, otherwise it will lock the row. Special Instructions: none ************************************* APAR NUMBER: IR51741 APAR Description: CM Object Server remote migration not correctly handling Network errors and leaving the Library Server in an inconsistent state. Fix Description: Modified UNIX calls for user information to correct for buffer problems Special Instructions: none ************************************* APAR NUMBER: IR51858 APAR Description: CM Object Server migrator candidate query failing with rc=-104 Fix Description: Syntax error removed that caused failure. Special Instructions: none ************************************* APAR NUMBER: IR51912 APAR Description: CMv71 Object Server Migrator stops or degrades due to an illegal operation as a result of Orphan Objects. Fix Description: Robust error processing was added to handle Orphan Objects. Special Instructions: none ************************************* APAR NUMBER: IR51966 APAR Description: CM Object Server Remote Migration is leaking memory when the source volume is TSM. Fix Description: Fixed memory allocation code to remove leak Special Instructions: none ************************************* APAR NUMBER: IR51967 APAR Description: Orphan objects in TSM change CM OS Migration performance to degrade. Fix Description: Dynamically detect Orphans objects in TSM and remove them from the migration candidate list. Plus fixed transaction processing to account for this abnormal behavior. Special Instructions: none ************************************* ************************************ Fixpack 710.17 ************************************ APAR NUMBER: IR51071 APAR Description: CM Object Server maintains a static internal table for outstanding mount request. This table currently has 26 rows and when filled, all I/O to CM Object Server is rejected. The conditions that cause this table to fill up are TSM mount requests for Optical Platters that take minutes to satisfy. Fix Description: The code was changed to allow this the Mount Table to be configurable so that a larger number of TSM Optical requests can be queued. Special Instructions: none ************************************* APAR NUMBER: IR51189 APAR Description: CM Object Server Remote migration fails when the file info cannot be found by the CM LS. As a result, CM OS Migration degrades significantly, since the same file or files show up in the migration list. Fix Description: The fix was to recognize this error condition, log it, and mark the problem file with a special flag so that subsequent migrations do not include this file. Special Instructions: none ************************************* APAR NUMBER: IR51383 APAR Description: The frnnlljb process is failing with a access violation trying to build a new static access module. Fix Description: For LS run on Window, this fix will use the dynamic query access for any new search path instead of building a new static access module. The current static access DLL's will remain unchanged for existing search path. This fix will prevent the LLsCreateDLLSource() function from failing in the future. Special Instructions: none ************************************* APAR NUMBER: IR51412 APAR Description: _IP2ListAccessList() API may crash due to memory exception if the number of access lists is high. Fix Description: In Ip2ListAccessList, another looping logic was changed so we check the number of detail data firs before && with other logic. This way we prevent from getting out of allocated boundary. Special Instructions: none ************************************* APAR NUMBER: IR50996 APAR Description: For a Search Transaction in the CMv71 Application: When the search attempts to return a result set of more than 3200 objects from the library server, they the client receives an FRN7002A error message. Fix Description: In CM71, there is limitation for the size of search result from a Search Transaction. If the size of the search result (in a single transaction) is greater than a total of 199 buffers (each sized 61440 bytes), then Library Server will return FRN7002A error message to the client. To avoid receiving this error, the user can limit the search criteria on the attribute values to prevent the search result response file exceeding the LS limitation. ************************************* APAR NUMBER: IR51349 APAR Description: MVA ViewSearch Error from eClient with large volume of MVA data. Return code 1 is issued by the CM 7.1 Library Server on a SimLibSearch call that searches an item with Multi Value Attributes, and has a ulMemListRequest = SIM_SEARCH_USER_SYSTEM_ATTR. An identical call with ulMemListRequest = SIM_SEARCH_MEMLIST works correctly. Fix Description: _Modify the dynamic query statement with multi-value attributes viewsearch to return first row back to client only to avoid the timeout and eventually terminate the child process while retriving the huge amount of data from LS database It is required that environment variable(FRNSINGLEATTRIBUTEVALUES=1) to be set on client machine when eClient is used to perform the search. Special Instructions: none ************************************* ************************************ Fixpack 710.16 ************************************ ************************************* APAR NUMBER: IR47149 Apar Description: CM AIX Servers not processing Dynamic Tracing correctly and not significant information is gathered. Fix Description: Fixed errors with CM AIX Servers (Library Server, Object Server) to process Dynamic Trace event and gather relevant information. Special Instructions: none ************************************* APAR NUMBER: IR50067 APAR Description: The History Log informational message FRN2259I can show that an object was removed from the wrong workbasket. Fix Description: Delete the Item which must match with the WorkbasketID received from FMClient. For Old LC API orders, the WorkBasketID is not required. Special Instructions: none ************************************* APAR NUMBER: IR50706 APAR Description: The History Log informational message FRN2259I can show that an object was removed from the wrong workbasket with the following scenario: . User1: Views WorkBasket 1 (WB1) User2: Views WB1, selects item 'A' and adds it WB2 User1: Selects item 'A' and removes it (notes: no refresh of the workbasket was done, and item 'A' is actually removed from WB2) User1: Does a search on all items to redisplay item 'A'. A view of the History Log shows item 'A' routed to WB2, but removed from WB1. . It is also possible that an item can be removed from a workbasket where the user does not have the authority to do so. The scenario is the same as above, except that User1 does NOT have authority to remove items from WB2. The actual remove call is successful. Fix Description: When this API (Ip2RemoveWorkbasketItem) is called, we need to forward the WB itemID to Library Server also, to allow LS to do privilege checking, and only perform the remove action if the workbasket itemID passed is correct. Special Instructions: none ************************************* APAR NUMBER: IR50776 APAR Description: Client application receives frn6772a (cannot complete the search) when attempting to perform a basic search against an index class which contains an attribute(s) defined as decimal (nullable). Fix Description: Fixed the internal logic error for the nullable decimal attribute. Special Instructions: none ************************************* APAR NUMBER: IR50794 APAR DESCRIPTION: The problem was caused during Basic Search on the IndexClass with multi-value attributes. The LS trace shown the following error msgs for Basic Search hitting MVA with the large result set spanning the response buffers. 1029 FRN fnc_data CLIENT LI (1.35.112.26) pid: 424, tid: 2708, cpid: -1 <2003-01-07 10:14:21.18 Type: ERRTRACE File: Liosock.c Line: 987 Func: IsoSockRecv RC: 10054 MsgCode: 9093> .FRN9093A: The sockets function recv returned error ECONNRESET. .. The SIMLIBGETTOCDATA in the FMTRACE.TXT did not complete. On the Client Application Window, the following error msg was displayed: FRN6772A: The system cannot complete the search. This error can occur with or without multi-value attributes. 2003-02-10-12.13.13.983690 Instance:FRN Node:000 PID:23410(frnxllch) Appid:none LIBRARY SERVER LL Probe:1 <2003-02-10 12:13:13.98 Type: ERRLOG File: frnpllbf.c Line: 224 RC: 7002 MsgCode: 7002> FRN7002A: The system could not allocate () memory of bytes in function . FIX DESCRIPTION: Fixed the logic error in the area of packing the query result set with multi-value attributes. Fixed the checking logic frnpllbf.c so that the total meory to be allocated is not restricted anymore. But still keep the checking for total number of buffer. Special Instructions: none ************************************* APAR NUMBER: IR50800 If a user, user1, checks out an item, but another user with higher privilege checks in the item on his behalf, user1 won't know this, and still thinks he has this item checked out. The check-in/ check-out info is stored in logon cache. So all check-in/check-out related API's, such as SimLibGetItemSnapshot, SimLibGetTOCData, SimLibCheckOutItem, and SimLibGetItemInfo, will all have wrong checked-out status. The only way to correct the cache data is to logoff first, then logon again to refresh data in cache. Fix Description: Foe the above-mentioned API's, code has been changed to ignore the check-out status from the cache, and always find the checked-out status from Library Server and use that data to return back to the caller of the API's. This way all the check-in/check-out status of any item is always consistent with the data in Library Server. Special Instructions: None. ************************************* APAR NUMBER: IR50829 APAR Description: When a user has lots of values for multi-value decimal type attributes, Library Client may calculate the row size of the returned table wrong. As a result, the 3354 error may occur. Fix Description: In Library Client code, when figuring out rowsize and number of rows, changed from using FIELDWIDTH to figure out each field length to use DmnFieldWidth in order to take care of multi-value attributes correctly. Special Instructions: A second fix has applied to Folder Manager. If a user uses a environmental variable, FRNSINGLEATTRIBUTEVALUES, then Library Server will only look up one row in the AVT and MVT tables when return attribute values back. In case of multi-value attributes, this will significantly speed up Library Server processing and returns the result back in sub-seconds or a few seconds, instead of minutes or hours (when huge amount of multiple values exist for multi-value attributes in MVT tables). One side effect is, for multi-value attributes, only one value will be returned, and this value may not be the same as the search criteria of the SimLibSearch() call earlier. However, this item should have one value in MVT tables that matches with the search criteria of the previously called SimLibSearch. ************************************* APAR NUMBER: IR50846 APAR Description: When a customer has a huge list of access lists, the Ip2ListAccessList() API call may fail and cause a memory exception. APAR fix description: In Ip2ListAccessList, the looping logic was changed so we check the number of detail data first before && with other logic. This way we prevent from getting out of allocated boundary. Special Instructions: none ************************************* APAR NUMBER: IR50908 APAR Description: When reindex is not successful, RC 7383 is returned. Customer recieves a 7383 when reindexing because of db2 deadlocks or SQL0911. Fix Description: Added the logic to check the lResult from dbUpdateWIPItemsClass(). The RC_ORDER_UNEXPECTED_SQL_ERROR with lsql_code: -911 will be returned if deadlock occurs. Special Instructions: none ************************************* APAR NUMBER: IR50960 Apar Description: CM Object Server Migrator error messages do not contain enougth information to resolve problem. For instance, Source file for migration and Target Storage Class info are needed to aid problem solving. Fix Description: Error message now contains detail information on Source File and Target Storage Class. Special Instructions: none ************************************* APAR NUMBER: IR50995 Apar Description: CM Object Server Migrator is not processing problem files (zero size, missing) gracefully. As a result, these files can continually show as Migration Candidates and generated excess error messages. Fix Description: Once the CM Object Server Migrator comes across these type of files (zero size, missing), it will update the DB with obj_status='E' so that subsequent migration queries will ignore these files. The CM Admin. can later go back and query for these files to determine the problem and appropriate action. Special Instructions: none ************************************* APAR Number: IR47609 APAR Description: The message text for an FRN9906A message is: FRN9906A %1 is not a valid migrator start time in the BASE_CONTROL table. This message is also used to indicate a problem with a replication start time. Fix Description: Message text should state: FRN9906A %1 is not a valid migrator or replicator start time in the BASE_CONTROL table. Special Instructions: None ************************************* APAR Number: IR48630 Apar Description: A French customer received 0403-057 from AIX after installing the library server when frnsetup.lib was run. The message that translates roughly to 'The environment for the library server is set' is posted by the file using ECHO. Apparently the format of the message caused the 0403-057. It has been confirmed some languages will encounter this message but it does not affect the result of the setup commands and can be ignored. This causes unattended installs to fail. Fix Description: Documentation. For some language, when running frnsetup.lib after install, users might see a 0403-057 error due to some syntax problem in an ECHO statement. Since this is a purely message and has nothing to do with the code, users can just ignore the error. Special Instruction: None ************************************ APAR Number: IR49760 APAR DESCRIPTION: Customer reported inaccuracy in the Retention period field of the Version 7 Release 1 Content Manager for Multiplatforms system administration online help. For the retention period, the user selects an interval for the number of days an object remains in the storage class. FIX DESCRIPTION: The online help incorrectly states that, if you change the retention interval, all objects already associated with the storage class receive the new value. The help should state that, if the retention interval is changed, only new objects receive the new value. The retention interval for already existing objects does not change. Special Instruction: None ************************************ ************************************ Fixpack 710.15 ************************************ ************************************* APAR NUMBER: IR50610 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: The rebind command does not rebind the CM's search results cleanup utility. As a result, after the customer applies the fixpack, he will not be able to run the cleanup utility. FIX DESCRIPTION: If the user needs to run the utility to clean up the old search result, the following bind command can be entered manually on the command prompt after fixpack 15 is applied. 'bind llpsrfcu.bnd datetime iso' ************************************* APAR NUMBER: IR48543 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: FRN7225A RESULTS IN THE STOPPAGE OF THE REPLICATOR Customer is recieving error message 7225 on a CM71 library server. As a result, the replicator treats this as a serious error and no longer continues to process items. The message is undocumented and provides little information on the symptom. FIX DESCRIPTION: Do not treat the 7255 as a fatal error. Log it but contnue the processing. ************************************* APAR NUMBER: IR50091A COMPONENT: LIBRARY SERVER APAR DESCRIPTION: FRN7449A is posted in the frndiag.log. Users may complain they can not log on. Users that are already logged on appear to function correctly. Further analysis shows frnheap starts growing at a fast rate. The FRN7449A appears to start occurring when frnheap reaches the limit of its definition (it may take more that an hour from the start of the heap growth to the first FRN7449A). FIX DESCRIPTION: The problem should only be occurred using Oracle DB. To bypass LLsSendQueryExprMessage() for Oracle DB. Then the FRNHEAP size can be stablized for Oracle. ************************************* APAR NUMBER: IR50153 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: The customer is seeing frn7449a error messages in the log file for the library server. It is the frnheap that has been exhausted. The client application was looping trying to connect to the Library server with an expired token. FIX DESCRIPTION: After LS removed the memory leak from the connect error logic code, the library server correctly closed the connection to the client with the FRN7055A message and returned code 7055 to the client application. ************************************* APAR NUMBER: IR50157 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: CM 7.1 LIBRARY SERVER MIGRATION UTILITY FAILS WHEN CREATING THE DLSAMPLE INDEX CLASS Content Manager 7.1.0 Fixpack 13 included a re-write of the Library Server database migration utilities. Running the utility against a database that does not have the DLSAMPLE index class defined produces the following errors in frnpmiut.log: ***** MODULE: checkSysIndexClass() LINE: 5519 RETURN CODE: -1 Action: Insert DLSAMPLE attribute into SBTATTRDEFS Detail: INSERT INTO SBTATTRDEFS VALUES (1, 448, 1, 50, 2048) Not available on AIX platform ***** FIX DESCRIPTION: Further inspection of the log shows that the CREATE TABLE for the DLSAMPLE index class specified the table name as AVT00000, and also tries to define the attribute names starting with ATTRIBUTE00000, then ATTRIBUTE00001, etc. Using these values does not work. SPECIAL INSTALLATIONS INSTRUCTIONS (if any): The migration utility frnxmiut.DB2.exe and the bind file frnomiut.bnd must be copied into the same subdirectory. The migration utility will dynamically bind the DB2 plan using the frnomiut.bnd as input file. ************************************* APAR NUMBER: IR50179 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: The problem was caused during Basic Search on the IndexClass with multi-value attributes. The LS trace shown the following error msgs for Basic Search hitting MVA with the large result set spanning the response buffers. 1029 FRN fnc_data CLIENT LI (1.35.112.26) pid: 424, tid: 2708, cpid: -1 <2003-01-07 10:14:21.18 Type: ERRTRACE File: Liosock.c Line: 987 Func: IsoSockRecv RC: 10054 MsgCode: 9093> .FRN9093A: The sockets function recv returned error ECONNRESET. .. The SIMLIBGETTOCDATA in the FMTRACE.TXT did not complete. On the Client Application Window, the following error msg was displayed: FRN6772A: The system cannot complete the search. FIX DESCRIPTION: Fixed the logic error in the area of packing the query result set with multi-value attributes. ************************** APAR NUMBER:IR47724 COMPONENT: Resource Manager APAR DESCRIPTION: The object server on first startup after migrating the code base to version 7.1 detects that the databases need to be migrated. If any problems are encountered during the table unload due to space problems with logs files, any restart of the CM migration fails since some tables already have been migrated. FIX DESCRIPTION: Enhanced error checking to ignore operations (column adds, etc.) that will normally fail, but are OK if the DB migration is restarted. ************************************* APAR NUMBER:IR48519 COMPONENT: Resource Manager APAR DESCRIPTION: Customer had manually changed REP_REPLICATETYPE to 'B' within the source CM71 object server BASE_REPLICATION table. Replication failed to take place and there was no error message posted to the frndiag.log that replication failed due to invalid rep_replicatetype. FIX DESCRIPTION: Problem fixed to report in the log an invalid replication status. Incorrect status values will be ignored. *************************************************** APAR NUMBER:IR49670 COMPONENT: Resource Manager APAR DESCRIPTION: AIX FILESPACE NOT MARKED FULL. FRN9816A DESTAGER ERROR FIX DESCRIPTION: Volume space calculations are updated to handle Volumes that are close to being full. *************************************************** APAR NUMBER: IR49853 COMPONENT: Resource Manager APAR DESCRIPTION: Validation utility receive sig 4 after processing approximately 85,000 records either in one validate request or a series of accumulated requests. Sig 4 is recorded in the frndiag.log and the validation process ends abruptly. FIX DESCRIPTION: Clean the memory leak problem in the validation utility code. Installation: Back up the original frnxlbsy(AIX), or frnnlbsy(Window) and apply the new one. *************************************************** APAR NUMBER: IR49907 COMPONENT: Resource Manager APAR DESCRIPTION: Problem with file permission on CM Object Server parts being modified unintentionally by other applications. As a result, the CM Object Server Purger will incorrectly delete files that should not be removed. FIX DESCRIPTION: Modified Object Server Purger to checks each file that is available for purging against the database to ensure this file was not incorrectly modifed to RW. If any file encounters this state, the file permissions are adjusted to the intended state and the file is removed from the purger list. *************************************************** APAR NUMBER: IR50203 COMPONENT: Resource Manager APAR DESCRIPTION: CM validation utility incorrectly creates a record in syncreport5 indicating that there is a createdate discrepancy between the records in sbtparts and base_objects. This appears to occur only for objects created after daylight savings time has taken affect. A comparison of the actual records from each of the tables shows that the createdate timestamp is the same. LOCAL FIX: manually validate the records reported. FIX DESCRIPTION: Make the code changes for day light saving on the timestamp. Installation: Back up the original frnxlbsy(AIX), frnnlbsy(Window), then apply the new one. ***************************************************** APAR NUMBER: IR50692 COMPONENT: Resource Manager APAR DESCRIPTION: An incorrect file size results in a destaging failure. FIX DESCRIPTION: The wrong format specifier was used for double data type. ************************************ Fixpack 710.14 ************************************ APAR NUMBER: IR49231 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: If a retrieve request from a Content Manager client takes a long time to be fulfilled (e.g. the retrieve involves a TSM retrieve from tape), it is possible that the Folder Manager on the client will time out. Once the object server begins to transfer the object to the toolkit daemon, the daemon complains that the file it is to place the object into is either null or unaccessible. The client frndiag.log will show the following errors: PID:68812(java) Appid:none CLIENT DA Probe:1 <2002-08-20 10:21:08.40 Type: ERRLOG File: dapobjst.c Line: 1359 Func: DmnOpenMapFile Action: pszFileName> List Manager gave Daemon null string for FileName 2002-08-20-10.21.08.557281 Instance:FRN Node:000 PID:68812(java) Appid:none CLIENT LI Probe:1 <2002-08-20 10:21:08.55 Type: ERRLOG File: lioapi.c Line: 3465 Func: IsoForceOutClient Action: DATA RC: 0> FRNODAOS closing communications with partner. 2002-08-20-10.21.18.619478 Instance:FRN Node:000 PID:68812(java) Appid:none CLIENT DA Probe:1 <2002-08-20 10:21:18.61 Type: ERRLOG File: dapobjst.c Line: 1902 Func: DmnRetrieveBlob RC: 9139> Daemon Object Server Thread=1286: Blob retrieved failed. Or the error may show up like this followed by the last two messages above PID:100236(java) Appid:none CLIENT DA Probe:1 <2002-08-20 12:12:57.14 Type: ERRLOG File: dapobjst.c Line: 1385 Func: DmnOpenMapFile Action: FrnSysOpen> Open error for file /home/mp2/frn/work/AHFMIMLP.TMP FIX DESCRIPTION: Validate the list manager element being used to store or retrieve. If the request had timed out then the element may be in use by another request. Validate the item id being processed is the same as that in the list element. ************************* APAR NUMBER: IR49232 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: Simultaneous retrieve of the same object from TSM into CM object server staging causes a file collision in the CM staging area . The second of the two retrieves will get an error during retrieve. This can result in a Folder Manager Timeout to the application. FIX DESCRIPTION: In the case where a collision is detected, the process detecting the collision will retry using the same file in staging, in effect waiting until the first retrieve request is satisfied and completes placing the object into staging. ************************* APAR NUMBER: IR49511 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: Simultaneous logons can potential cause a race condition when initializing the list manager. If the problem occurs, the second user will get a 9126 LSTERR_INSUFFICIENT_MEMORY FIX DESCRIPTION: When a second process or thread toolkit application attempts to initialize the list manager component, it will detect that initialization is in progress by the first. The second process will sleep for 2 seconds and then attempt to attach to the already initialized list manager. If this attach still fails a more meaningful error code is returned. LST_ERR_INITIALIZATION_FAILED 9101 ************************* APAR NUMBER: IR49630 COMPONENT: Library Server APAR DESCRIPTION: There are cases in the library server where it is not returning sqlcodes to the client. This is masking db2 sqlcode that the client needs. FIX DESCRIPTION: The code will now pass back the sqlcode in the lsqlcode fields of the response block. ************************* APAR NUMBER: IR48747 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: When ddo.add() or ddo.update() is inside a pending transaction, users cannot log into the system. The connect methods will just block until that transaction is complete. FIX DESCRIPTION: Fixed ************************* APAR NUMBER: IR49491 COMPONENT: 5648D0431 / Library Server APAR DESCRIPTION: LIBRARY SERVER child processes terminating unexpectly with no error messages in FRNDIAG.LOG or ERRPT FIX DESCRIPTION: Fixed. ************************* APAR NUMBER: IR49531 COMPONENT: CS C API Toolkit APAR DESCRIPTION: SIMLIBDELETEITEM API does not correctly map the SQLCODE -911 returned by the CM Library. Swveral other API's have the same problem. FIX DESCRIPTION: Changed the code of SimLibDeleteItem and other API's to make sure when Library Server returns a bad return code, the SQL code will always be surfaced as ulExtReason in the RCSTRUCT returned data structure. ************************* APAR NUMBER: IR49223 COMPONENT: Library Server APAR DESCRIPTION: STAGING AREA FILLS UP, FRNCMD STATUS COMMAND SHOWS A LOT OF SPACE LEFT IN STAGING WHEN THE FILE SYSTEM IS ACTUALLY FULL 'frncmd (object server) status' shows only about 25% used. Since the purger is set to start when the staging area is 80% full, it never runs. FIX DESCRIPTION: Set the purger start / stop percentages very low (e.g.: 20 / 10 ) ************************* APAR NUMBER: IR49740 COMPONENT: Library Server APAR DESCRIPTION: When the Library Server determines there is a need to cancel a transaction, as when a communication error occurs between the client and LS, or the LS receives a timeout waiting for requests from an active client transaction, the library server would roll back the DB2 transaction, but fail to send a rollback or an error to the object server. This would potentially leave work in the transaction uncommitted when the conversation was terminated as if no error occurred. FIX DESCRIPTION: The Library Server failed to send endTrans and then closes the connection cleanly. This was leaving uncommitted data. OS should rollback if commit not received. ************************* APAR NUMBER: IR49275 COMPONENT: Library Server APAR DESCRIPTION: Application program hangs on a retrieve request to CM. The object requested is never sent to the application. The object is also not found in the staging area. A trace of the Object Server shows that the object is in TSM, and that TSM sent back a return code '2' (object not found - DSM_RC_ABORT_NO_MATCH ). FIX DESCRIPTION: Fixed the hang. ************************* APAR NUMBER: IR49825 COMPONENT: Library Server APAR DESCRIPTION: Malloc splint result in a lost FRNXLLCH child process FIX DESCRIPTION: When the LS gets a fatal error (in this case, a 911) and takes the error path, it flags the OS involved in the transaction as inactive and not updated. The later flag was not being reset properly. The LS now sets that flag appropriately. Also, the isolator component avoids using its transmit buffer, when it determines that its socket handle is invalid and will return ISOERR_BAD_PARTNER_HANDLE = 9001. Additional trace was added to the isolator to trace its memory allocation and free of its transmit buffers. Finally, the LS will no consider an endclient failure to the OS as FATAL. The transaction can rollforward. ************************************ Fixpack 710.13 ************************************ APAR NUMBER: IR48747 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: When ddo.add() or ddo.update() is inside a pending transaction, users cannot log into the system. The connect methods will just block until that transaction is complete. This fix was initially provided in fixpack10. Fix Description: Moved the checkout table queries to a different package and used the UR isolation level on that package. ************************* APAR NUMBER: IR47881 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: The FRNNMIUT.exe fails to commplete. And examination of the failure show that we are looking FRN$SEARCH indexclass in the SBTNLSKEYWORDS table. EXEC SQL DECLARE SearchLangCursor CURSOR FOR SELECT LANGUAGECODE FROM SBTNLSKEYWORDS WHERE KEYWORDSTRING LIKE '%FRN$SEARCH' AND KEYWORDCLASSFI = 2 FOR READ ONLY; This fails with a SQLCODE of 100, row not found. FIX DESCRIPTION: Updated the FRNNMUIT.exe migration utility. Refer to install instructions. ************************* APAR NUMBER: IR49024 COMPONENT: TOOLKIT APAR DESCRIPTION: During transaction processing to load new objects into Content Manager, locks are obtained and held on the SBTCHECKEDOUT table. If the transaction runs for a long time deadlocks can occur between the load program and normal retrieve processing. Since the load is done with a transaction, the locks on SBTCHECKEDOUT are not necessary for the load program. FIX DESCRIPTION: If we are in a transaction, then we don't need to do check in/out while doing updates. This includes APIs like: SimLibOpenItemAttr, SimLibCloseAttr, SimLibCatalogObject, SimLibCreateObject, SimLibStoreObject, and SimLibOpenObject. ************************* APAR NUMBER: IR46389 COMPONENT: INSTALL APAR DESCRIPTION: The frncrtdb.lib.sh shell script calls frncrtd1.sh. If on a normal system, frncrtd1.sh calls frnxinlc.db2 to create the db tables. However, if on a DB2 EEE system, frncrtd1.sh calls frnxinle.db2. The frnxinle.db2 shell script has two problems: 1. The command to create an index on SBTFOREIGNPARTSNMS has a bad column name in the command. This causes a rc=8930, and the db creation process halts. 2. The frnxinle.db2 shell does not seem to be updated. It creates a CM v6.1 database (it should create a v7.1 database). The new tables are missing (ie SBTCOPIES) . When the bind runs, "table not found" error messages are received. FIX DESCRIPTION: The AIX script frnxinle.db2 has been updated with the new tables. ************************************ Fixpack 710.12 ************************************ APAR NUMBER: IR48765 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: On CM71 AIX and NT, customer is unable to delete a zero-length blob (DDO) with content class set to DK_DL_CC_UNKOWN using the EIP71 java APIs. The error code returned is FRN6056, ExtRC=7736 FIX DESCRIPTION: Previously, LS detects the blob has zero length and returns this error. Change has been made to return OK since Obj Server can handle deleting zero length object. ************************* APAR NUMBER: IR47095 COMPONENT: COMMON SERVICES APAR DESCRIPTION: In CM7.1, if the network configuration has been changed to invalidate the FRNOLINT.TBL, for example the hostname of a server is changed, then the network table generation utility will error with a "No servers read ERROR(9019): Network Table file syntax invalid, missing or misspelled". The erroneous FRNOLINT.TBL cannot be edited with the network table generation utility. FIX DESCRIPTION: Ensure that an invalid hostname can be read in, and altered via the network table utility. ************************* APAR NUMBER: IR48914 COMPONENT: OBJECT SERVER APAR DESCRIPTION: TSM Retrieve buffer size is too small. FIX DESCRIPTION: TSM Buffer size on Retrieve was increased to 32K. ************************************ Fixpack 710.11 ************************************ APAR NUMBER: IR47816 COMPONENT: OBJECT SERVER APAR DESCRIPTION: CM Object Server will not start if database name non-alphabetic character. FIX DESCRIPTION: Fixed Object Database algorithm to handle special characters. ************************* APAR NUMBER: IR48981 COMPONENT: COMMON SERVICES APAR DESCRIPTION: On AIX When using the frncmd or frnstop commands, a core dump is genereated as the command is exited. FIX DESCRIPTION: Upon shutdown, check if variables are initialized before using them to perform cleanup exit processing. ************************* APAR NUMBER: IR48665 COMPONENT: TOOLKIT APAR Description: In the use of the SimLibCopyObject a local copy of the object is stored at the Client workstation in the working directory for the client application. This file is not removed from this directory after use. This eventually fills the file system and causes the Dr. Watson error. Fix Description: At the end of SimLibCopyObject, remove the temporary file in FRNROOT\work directory. ************************* APAR NUMBER: IR47981 COMPONENT: OBJECT SERVER APAR DESCRIPTION: CM Object Server intermitently fails to initialize on Oracle with certain TSM volume names. FIX DESCRIPTION: Fixed algorithm that parsed the volume information to strip invalid information. ************************************ Fixpack 710.10 ************************************ APAR NUMBER: IR46942 COMPONENT: LIBRARY SERVER APAR DESCRIPTION: The Library Server intermittently hangs during shutdown. FIX DESCRIPTION: Shutdown logic changed to avoid the failing code. Removed calls to handshake between components and avoid looping caused by memory faults. ************************* APAR NUMBER: IR47443 COMPONENT: OBJECT SERVER APAR DESCRIPTION: LBSDESTAGEOBJECT MSGCODE 9883 DOES NOT EXPOSE THE RETURN CODE DURING TRACE FIX DESCRIPTION: Program was changed to expose the ERRNO information in the Destager Trace. ************************* APAR NUMBER: IR48435 COMPONENT: OBJECT SERVER APAR DESCRIPTION: In the move of the retrieved object from the Object server to the LAN Cache server the opening of the object file fails. The order being processed is LbsProcessMoveFromOrder which fails with the 9922 message. In testing it was discovered that a sharing violation was received for the open request. On examining the trace we found the two different open requests. The initial open for the retrieve request and the second request for the move to the LAN cache server are not using compatible opens so that both processes can have access to the file. This issue is only apparent if the object being retrieved is large. This customer found it at around 900KB files. FIX DESCRIPTION: The file open attributes during a "move from" internal order were changed to an explicit shared_read. This change impacts only NT. ************************ APAR NUMBER: IR48544 COMPONENT: 5648D0411 / OBJECT SERVER APAR DESCRIPTION: Problem was caused by the CMv61 command utility application not having a large enough buffer to contain the output of the GECOS text for the AIX user. The resulting error was the AIX User information could not be read and the command utility failed to operate because of the authentation failure. FIX DESCRIPTION: Program was changed to use a larger buffer to handle the GECOS information. ************************** APAR NUMBER: IR48676 COMPONENT: COMMON SERVICES APAR DESCRIPTION: After installation of CM 7.1.0 Fixpack 9, various errors/traps and/or Access Violations can occur if FRNTRACELEV is not set to '0'. The problem shows itself during application initialization or the first time store or retrieve object is performed. On the Library Server, an Application Exception is seen in FRNNLLCH.EXE during object import. The error window has text like: "The instruction at "0x005a052a" referenced memory at "0x00000069". The memory could not be "read". Drwtsn32 shows the error to be an Access Violation ( 0xc0000005 ). The child process associated with this error is terminated. The LS could eventually stop responding if all child processes experience this error. On the production client (vic.exe), the problems shows up as no servers listed in the drop down listbox in the logon panel. On customer applications, Access Violations (0xc0000005) will occur. FIX DESCRIPTION: Removed use of uninitialized variable pointers for specific logging instruction. **************************************** Fixpack 710.9 **************************************** APAR NUMBER: IR47761 COMPONENT: Object Server APAR DESCRIPTION: The customer was saving an update to an object but received RC 6056 ExtRC 9924 through the custom client. The Library Server posted FRN7030 with an RC 9924. The Object Server logged no errors. A subsequent retrieve fails with the error analysis exposing the fact that the file on the object server is missing. The BASE_OBJECTS row for the missing file indicates it had been destaged. A trace of the object server shows in the replacement of the file, the discard order fails due to a database error and the retry also fails. Development believes the transaction processing is allowing the deleting of the original file even though the update failed. FIX DESCRIPTION: The old object will not be deleted until destager runs. ************************* APAR NUMBER: IR47817 COMPONENT: Common Services APAR DESCRIPTION: AIX applications would fail to logon with 9045 return code. FIX DESCRIPTION: The initialization of the toolkit was not threadsafe in reading in the network table. A previous APAR addressed this problem, but broke some AIX applications. Hence, threadsafe initialization was modified for the AIX CM Toolkit, such that it didn't rely on C++ constructors which failed to run in some circumstances. SPECIAL INSTALLATION INSTRUCTIONS: Any workaround, in setting FRNSINGLEPROCESSCLIENT=0 or 1, should be removed. ************************* APAR NUMBER: IR48102 COMPONENT: Content Server APAR DESCRIPTION: IP2LSTUSERS and IP2LSTUSERGROUPS limits the display of users to 10000. FIX DESCRIPTION: Increased Ip2ListUsers limit from 10,000 to 30,000. ************************* APAR NUMBER: IR48114 COMPONENT: Library Server APAR DESCRIPTION: Search on an index class with option ATTRONLY results in a termination of a child process at the library server. No coredump created. FIX DESCRIPTION: Fixed the mismatching structure definition that was causing memory corruption. ************************* **************************************** Fixpack 710.8 **************************************** APAR NUMBER: IR47140 COMPONENT: 5648D0422 / LIBRARY SERVER PMR NUMBER: 44932,044,866 APAR DESCRIPTION: This problem shows up if a decimal keyfield has been defined and is included in an indexclass. A search on the this indexclass caused the library server child process, FRNNLLCH.exe, to take the application error reported by Windows. FIX DESCRIPTION: Fixed how we packed a single valued decimal field. ************************* APAR NUMBER: IR47585 COMPONENT: 5648D0431 / COMMON SERVICES PMR NUMBER: 47516,075,724 APAR DESCRIPTION: The frnxdamn process is not starting as a separate process in the AIX client after the install of CM 7.1.0 Fixpack 3 FIX DESCRIPTION: Altered change made for Windows Toolkit so that it doesn't apply to AIX. Hence, the AIX Toolkit will behave as it did prior to Fixpack 3 in that frnxdamn will start as a process. SPECIAL INSTALLATION INSTRUCTIONS: Any workaround, e.g. setting FRNSINGLEPROCESSCLIENT=0, should be removed ************************* APAR NUMBER: IR47882 COMPONENT: 5648D0421 / LIBRARY SERVER APAR DESCRIPTION: Performing an update operation on a object within a transaction puts a lock on the checkout table for that item. This blocks other queries for that item. Fix Description: Moved the checkout query to a different package with a different isolation level. Special Install Instructions: There is a new bind file (frnollq6.bnd) that is created by this fix. Make sure the new bind file is copied to the BIND directory of the CM install directory. ************************* APAR NUMBER: IR48012 COMPONENT: 5648D0452 / CS CAPI TOOLKIT PMR NUMBER: 73679,500,000 APAR DESCRIPTION: Using the standard Content Manager client (vic.exe), message 'FRN6531A You cannot open the history log' is seen by users on some machines when trying to open the History Log. The same user (using the same userid) can successfully retrieve the history log on other machines. Folder Manager trace shows an Ip2ListHistory call with ulRC = 6288 (A library client call within a folder manager function was not successful), and a ulExtRC = 9107 (The element handle is not valid). FIX DESCRIPTION: In retrieving the history log with event type parts, prior to the LibRetrieveItemPart() call, added two more calls, LibAllocateItemPart() and LibNameItemPart(), to associate the part into an entry in the element table. Later on when the retrieve order is launched, the Daemon can uniquely identify the correct entry in the element table that corresponds to the event part being retrieved, so the retrieve will be successful. **************************************** Fixpack 710.7 **************************************** APAR NUMBER: IR47838 COMPONENT: 5648D0452 APAR DESCRIPTION: In Folder Manager simLibLogon code, number of workbaskets and workflow are 3000 maximum separately. This is too restrictive for customers. A higher limit, like 30000 is needed for both. FIX DESCRIPTION: Increased both limits to 30000 from 3000. ******************* APAR NUMBER: IR47744 COMPONENT: Library Server APAR DESCRIPTION: Content Manager 7.1 for AIX supports multiple instances of the library server. There exists no documentation on proper setup of such a configuration. FIX DESCRIPTION: CM v7.1 will allow you to run two instances of the same version of the library server on the same machine. You will install the library server code as you would as normal. You will need to create two different databases for the library server. The database names must be different. You may have the two library server databases in the same instances or different instances. You will need to create an entry in the network table for both of these servers in the network table. Use the network table generation utility to accomplish this. The index class and the view search dlls for each server will be placed in a different directory. The location of these dlls will be as follows: [FRNLOCAL]\[DB Name]\dll. On AIX, you will have to define two different users to administer each server. You must also run frnconfig.lib and frnstart.lib commands as normal for each of the servers. ******************* APAR NUMBER: IR47578 COMPONENT: Library Server APAR DESCRIPTION: When processing a forced log-off order from a client, there is a chance that the child process gets signal 11 and causes the library server to hang. FIX DESCRIPTION: During processing a forced log-off order, the child process will free a log-on instance block. There is a chance that it later refers to that block again and gets a segmentation violation. Changes have been made to make sure that it never refers to that freed memory block. ******************* APAR NUMBER: IR47652 COMPONENT: 5648D0452 / Toolkit APAR DESCRIPTION: If the default install directory of the CM toolkit or client is in a directory which contains a space, then in firewall cases, specifying FRNSINGLEPROCESSCLIENT=0 will cause the application to fail to logon to the CM library server. FIX DESCRIPTION: Remove startup dependency on path. ******************* APAR NUMBER: IR47261 COMPONENT: Library Server APAR DESCRIPTION: In those locales where the comma is defined as the decimal separator, a Simlibsearch results in an sql -104, an unexpected token was found. The search does not complete. FIX DESCRIPTION: Changed to code to make sure the decimal token is always "." irregardless of the locale settings. ******************* APAR NUMBER: IR46737 COMPONENT: 5648D0452 (C API Toolkit Win) APAR DESCRIPTION: Custom application using EIP api's to aix cm712 toolkit and library allows an user defined with viewonly privileges to update and delete library database records. FIX DESCRIPTION: We removed privilege bits 110, 111, 112, 113, from OIM_SA_USER, and added privilege bits 56, 57, 58 to OIM_SA_WORKBASKET and OIM_SA_WORKFLOW, and added privilege bits 110, 111, 112, 113 to OIM_CHANGE_INDEX_CLASS. REFERENCE DOCUMENTATION: Users have to apply this APAR fix first, then delete all existing privileges, and redefine them using sys admin before they get the benefit of this APAR fix. **************************************** Fixpack 710.6 included no AIX fixes **************************************** **************************************** Fixpack 710.5 **************************************** APAR NUMBER: IR46801 / AIX COMPONENT: OBJECT SERVER / 5648D0411 PMR NUMBER: 51946,180,000 APAR DESCRIPTION: If the "gecos" information field is updated in the /etc/passwd/ for the object server osadmin account, status commands to the object server fail with errors. . Issuing the following command to the object server: su - osadmin -c frncmd OSMTXH status . Outputs the errors: FRN2021I: Session Established at Fri May 4 11:28:12 EDT 2001 FRN8095A: Error. An incorrect parameter was entered. FRN8091A: Invalid user. FIX DESCRIPTION: Fixed problem with user descriptions which caused the cmd utility to fail. ******************* APAR NUMBER: IR44982 / AIX & Windows COMPONENT: LIBRARY SERVER / 5648D0422 PMR NUMBER: 17111,519,000 APAR DESCRIPTION: An application written using the OO-API may get the following error message when trying to work with a video object: Error text: DGL0423A: The specified part object does not exist SimLibQueryObject [FRN=6056, ExtRC=9924] In this example, the getCreatedTimestamp() method was being used, which eventually gets turned into a SimLibQueryObject call, which does not work correctly for video objects. FIX DESCRIPTION: As part of retrieving an object, file manager issues a query object API. This API sends an order to the object server for SMS info. A video object does not have any SMS info so this APIreturns an error and retrieve fails. ******************* APAR NUMBER: IR47426 / AIX & Windows COMPONENT: CS CAPI TOOLKIT / 5648D0451 PMR NUMBER: 48241,000,834 APAR DESCRIPTION: The following error message is seen when initializing a multi-threaded CM application program: 2001-12-04-12.07.59.409168 Instance:FRN Node:000 PID:20164(java) Appid:none CLIENT LI Probe:1 <2001-12-04 12:07:59.40 Type: ERRLOG File: lioiso.c Line: 1372 Func: IsoSocketsInit Action: SymInstallSymbol RC: 1 MsgCode 9066> FRN9066A: A startup problem occurred in the FRNOLIIS program. Return code = 1 The return code '1' indicates that the symbol is already installed. FIX DESCRIPTION: Correct daemon and IsoInitialization processing to avoid listservers threading problem. **************************************** Fixpack 710.4 **************************************** APAR NUMBER: IR45075 / AIX & Windows COMPONENT: Object Server PMR NUMBER: 27196,379,000 APAR DESCRIPTION: CMv71 Object Server reported a sql return code from DB2 as an application problem and did not report a severe error FIX DESCRIPTION: CMv71Object Server now reports the sql error as severe. REFERENCE DOCUMENTATION: None ******************* APAR NUMBER: IR46768 / AIX & Windows COMPONENT: dbmigration APAR DESCRIPTION: The database migration utility does not take into conisderation non-english languages with regards to attribute names. There are several different error messages the customer will receive depending on the error condition. Some examples include: Error in module 'ModifySBTFOREIGNPARTSNMS()', line #681: Error in module 'ModifyFRNSearchIC()', line #702: Affects: Windows and AIX platforms, db2 and oracle migration utilities for Content Manager 7.1 FIX DESCRIPTION: The migration utility has been modified to account for a non-standard CM tablespace in Oracle and the use of a LIKEclause instead of equality check for NLS keyword names (to accommodate for the * prefix). Also, SQL return code isalways displayed in case of error for better diagnosis of future problems. SPECIAL INSTALLATION INSTRUCTIONS: As always, the migration utility must be copied from the CD-ROM into a local temporary directory with read/writeprivileges. The directory into which it is copied must reside on the system with the Library Server to be migrated.For DB2, the bind file must also be copied into the same directory as the migration utility executable. REFERENCE DOCUMENTATION: Planning and Install Guide, migration section for general migration instructions. ******************* APAR NUMBER: IR45069 / AIX & Windows COMPONENT: Object Server PMR NUMBER: 27196,379,000 APAR DESCRIPTION: Customer was receiving FRN9006A, FRN9927A and FRN9919A in frndiag.log on retrieve. The trace uncovered that the trigger for the FRN9006A was 'CreateFileMapping RC: 8>'. The actual cause was that the file returned from ADSM was smaller than what was expected. FIX DESCRIPTION: Utilization of the supported ip2toccount api (as used by the vi client application) to determine the count of items in a workbasket results in a lenghthy response time when the number of items and workbaskets is large. SPECIAL INSTALLATION INSTRUCTIONS: 1. Rename \frenroot\dll\frnnfios.dll to something else. 2. Copy the new frnnfios.dll to the \frnroot\dll directory. 3. Bring up your application and try again. ******************* APAR NUMBER: IR45001 / AIX & Windows COMPONENT: C API Toolkit APAR DESCRIPTION: A CM user is created with privilege bit 15 cleared, which should allow filtering of search results to exclude all documents that the user does not have access to. However, when a search is done, a folder manager trace shows that it fails on a SimLibGetTOCData call with RC = 6238. The client (vic.exe) does not have the capability of filtering this way. This APAR is being opened per development to fix this problem in the folder manager. FIX DESCRIPTION: Change the SimLibGetTOCData source code to continue processing (instead ofreturning error return code 6238) when there is no information returned from Library Server for aspecific itemID that this user has no authority to access. In this case, the returned SNAPSHOTSTRUCTdata structure for this itemID is empty other than the itemID itself. ******************* APAR NUMBER: IR42801 / AIX COMPONENT: Library Server PMR NUMBER: 83284,422,000 APAR DESCRIPTION: FRNXMILE gets SQLCODE 204 due to root being used as the high level qualifier for the tables. FIX DESCRIPTION: ROOT userid hardcoded in Korn Shell script **************************************** Fixpack 710.3 **************************************** APAR NUMBER: IR43090 COMPONENT: OS PMR NUMBER: 64570,519,000 APAR DESCRIPTION: After the migrator migrates an objects into a storage class, it will set the OBJ_ACCTIONDATE (the next migration date) to a new date which matches the retention period, for example, if the retention period of the destination storage class is set to 60 day, then after migrating into this storage class, the OBJ_ACTIONDATE of an object will be set to 60 days from the current date. This means, during the next 59 days, the migrator will bypass this object when selecting candidates for migration. If the retention period has been set as "FOREVER", than after the migration, the OBJ_ACTIONDATE will be set to 12/31/9999, which means that the migrator will never consider it again as a potential candidate for migration. During every migration cycle, the migrator will "look" at any object with OBJ_ACTIONDATE set to the current day or early, and decide is the object due to migrate. Currently, unless the user set the retention period of a storage class to "FOREVER" or to a date before the year 2038 (some 13000+ days from now), the OBJ_ACTIONDATE of the migrated object will be set incorrectly to the next day. As more and more objects been migrated to the storage class with the wrong action date set, the migrator will waste most of the migration cycle time looking for objects that need not be migrated at all. Eventually, in some extreme case, all the migration cycle time are wasted, and there is no time left for any real object migration in the migration cycle. The migrator is using the ANSI C's time() function to compare two dates, but the time() function always return a negative value for any date after Jan 18, 2038, and this negative time will confuse the migrator and make it to set the next migration date incorrectly. FIX DESCRIPTION: The fix put in a check for the negative time returned from the time() function, and will set the next action date to 12/31/9999 in this case. So, if the user set the retention date of a storage class in the migration policy after Jan, 2038 (for example, from 13000+ to 32000 days) it will be treated as FOREVER. ************************* APAR NUMBER: IR44885 COMPONENT: LS / 5648D0431 APAR DESCRIPTION: If configuring a CM system such that the CM application is outside a firewall (i.e. a network table entry for the client daemon was added to the network table) searches to the library server, and perhaps even logon to the library server will fail with an out of memory condition. There should also appear rouge frnndamn.exe processes after each stop and restart of the application. This fix requires the user to specify the FRNSINGLEPROCESSCLIENT environment variable when configured to use a firewall. In addition to this fix, the application writer should consider making sure that all retrieval of objects are placed into files rather than into memory, thus avoiding memory full conditions. Error number 9126 (FRN9126A) will be logged in the client frndiag.log as a result of this failure. FIX DESCRIPTION: CM disconnect/exit processing was changed to use exit lists rather than destructors, changing the order in which CM would cleanup. ************************* APAR NUMBER: IR45362 COMPONENT: OS / 5648D0411 PMR NUMBER: 50501,032,724 APAR DESCRIPTION: Object server reports FRN9919A, FRN9883A on retrieval of an object, followed by an abnormal termination of the FRNXLBPR process, FRNsysSignalHandler, this process received a signal 4. Problem appears to be similar to symptom described in VI v24 apar IR42465. FIX DESCRIPTION: fix lbsretriev.c lbsmigra.c, and lbscomm.c. Filter out the zero size file. ************************* APAR NUMBER: IR45703 COMPONENT: TK / 5648D0452 APAR DESCRIPTION: AIX frnxdamn process will end with an exception when running with Multithreaded applications like Dynamic Page Builder and EIP. The problem is no limited to these applications and could occur with any AIX application. FIX DESCRIPTION: A zero file handle(0) is considered valid during object store and retrieve within the object server on AIX. ************************* APAR NUMBER: IR46071 COMPONENT: LS PMR NUMBER: 03486 APAR DESCRIPTION: In the 'Selected' pull down in the sysadmin are three utilities (Optimize, History, Expired time). Using the Optimize utility for testing, once it is set, it can not be unlocked. It appears to take the update but a subsequent logon displays the same scheduled status. FIX DESCRIPTION: Provided a new pgm to delete the utility schedules. SPECIAL INSTALLATION INSTRUCTIONS: For DB2: Issue the following commands to bind the package: db2 connect to db2 bind utpschd.bnd ************************* APAR NUMBER: IR46471 / Windows & AIX COMPONENT: 5648D0452 PMR NUMBER: PMR 46692,019 APAR DESCRIPTION: The nhttp.exe application on the Notes server was crashing causing the Domino server to fail. The DLL involved is the FRNNDAPI.dll. This is to handle CM crash problem. This problem can occur on CM 6.1 and CM 7.1.This problem can occur on CM 6.1 and CM 7.1. FIX DESCRIPTION: In DmnGetResponse function, added semophore protection before the SymDeleteSymbol() function call, and release the semophore after the call. This way there will be no interruption between different threads on the symbol table. ************************* APAR NUMBER: IR46771 COMPONENT: LS PMR NUMBER: 36410,075,724 APAR DESCRIPTION: The following message is posted at the object server: <2001-06-27 10:43:43.10 Type: ERRLOG File: lbsdb.c Line: 8984 Func: LbsUpdateTargetObjectRowReplace RC: -911 MsgCode: 9861> FRN9861A:Database operation type UPDATE for the LAN-based object server finished. Return code = -911. <2001-06-27 10:43:43.30 Type: ERRLOG File: lbsdb.c Line: 8798 Func: LbsFetchSourceObjectRowReplace RC: 100 MsgCode: 9861> FRN9861A: Database operation type FETCH SOURCE for the LAN-based object server finished. Return code = 100. FIX DESCRIPTION: Removed the deadlock retry logic from the library server. It was causing problems with the transaction commit scope on the object server. ************************* APAR NUMBER: IR46843 COMPONENT: LS / 5648D0432 APAR DESCRIPTION: If configuring a CM system such that the CM application is outside a firewall (i.e. a network table entry for the client daemon was added to the network table) searches to the library server, and perhaps even logon to the library server will fail with an out of memory condition. There should also appear rouge frnndamn.exe processes after each stop and restart of the application. This fix requires the user to specify the FRNSINGLEPROCESSCLIENT environment variable when configured to use a firewall. In addition to this fix, the application writer should consider making sure that all retrieval of objects are placed into files rather than into memory, thus avoiding memory full conditions. Error number 9126 (FRN9126A) will be logged in the client frndiag.log as a result of this failure. FIX DESCRIPTION: Removed the attempt to autodetect a firewall configuration and autostart the daemon process. The adminstrator is required to set FRNSINGLEPROCESSCLIENT. SPECIAL INSTALLATION INSTRUCTIONS: In a firewall configured environment: Determine if there is to be more than one application process on the client workstation. 1) If multiple applications are to be run on the workstation specify in the system environment variables or setup scripts: FRNSINGLEPROCESSCLIENT=0 and increase FRNHEAPSIZE to 2048 or higher. If out of memory conditions still occur, preferably, the application writer should, consider making sure that all object retrievals are placed into files rather than into memory or increase the FRNHEAPSIZE setting, thus avoiding memory full conditions. 2) Otherwise, when only one application process exists on the workstation, specify FRNSINGLEPROCESSCLIENT=1 and leave the default FRNHEAPSIZE setting. ************************* APAR NUMBER: IR46900 COMPONENT: TK PMR NUMBER: 14826,010,618 APAR DESCRIPTION: The Library Client API set is a published, supported interface to the Content Manager servers. To get the correct column width regardless of data type (IE: when using the LibGetQueryReplyInfo call), DmnFieldWidth found in FRNNDAPI.LIB must be used IF any multi-value attributes are used and multiple values are stored. FRNNDAPI.LIB and the associated header file were not shipped with CM 7.1. FIX DESCRIPTION: Update the file groups/filesets to install the missing file. ************************* APAR NUMBER: IR46902 COMPONENT: OS / 5648D0411 PMR NUMBER: 46832,519,000 APAR DESCRIPTION: CM71 VALIDATION REPORT RESTORE CAPABILITY FOR TYPE 7 DISREPANCY RESULTS IN A SIGNAL 11. CM71 aix validation utility correctly identifies a type 7 discrepancy, however when the restore function is selected, the process ends with a signal 11 on the replication request. FIX DESCRIPTION: Change replicator program, new frnxrepli created.. Customer need to apply this new module. ************************* APAR NUMBER: IR45810 / AIX COMPONENT: 5648D0451 APAR DESCRIPTION: Turning off CM client default signal handler FIX DESCRIPTION: Java applications using the CM client code can now turn off the CM signal handler in order to determine their own application problems. Before running the Java application, a user should export FRNSIG=0 to turn off signal handler in CM client code. If this environment variable is not set or is set to a different value, the CM signal handler is on. Note that the detection of this environment variable occurs at application initialization time. Therefore, once the application is started, changing the value of the environment variable does not turn on or off the CM signal handler. ************************* APAR NUMBER: IR46505 / AIX COMPONENT: dbmigration APAR DESCRIPTION: In some Oracle environments, language code is not correctly stored internally after being read from an Oracle table. Problem is unpredictable and does not surface on all Oracle installations. The error has occurred in some environments and not in others. Customer will encounter an error during the execution of the migration utility for an Oracle system. Error will be similar to, "Error in module ModifyFRNSearchIC(), line ..." Migration will halt. FIX DESCRIPTION: Internal character array used to store language code expanded well beyond the 3-character language code read from Oracle table. ************************* APAR NUMBER: IR46602 / AIX COMPONENT: LS / 5648D0421 PMR NUMBER: 42028,010,678 APAR DESCRIPTION: Customer used frncrtdb.lib and frncrtdb.obj to create the Library and Object server databases CM_LIB and CM_OBJ. When they start the server, they get a message: "The server cannot be validated". If they create the database without the underscore in the name (i.e. CMLIB and CMOBJ) then they can start the servers with no problem. CM_LIB AND CM_OBJ are valid database names as DB2 UDB does allow some non-alphabetic and non-numeric characters to be used in database names. See the DB2 Quick Beginnings manual, Appendix C, Naming Rules for details. FIX DESCRIPTION: Removed the db name checking in the library server. Let it be enforced by db2 at the database creation time. **************************************** Fixpack 710.2 **************************************** APAR NUMBER: IR46504 / AIX COMPONENT: Object Server PMR NUMBER: 09795,519,000 APAR DESCRIPTION: John Rodriquez's group encountered a problem with the bind utility on the object server. This was a VI 2.4.1/UDB5.2 on AIX 433 to CM 7.1/UDB7.1 on AIX 433 upgrade. Apparently the bind was triggered at object server initialization but it failed with: <2001-06-24 00:13:36.72 Type: ERRLOG File: lbscmn.c Line: 9191 Func: LbsInitializeDBManager2 Action: LbsBindDB RC: 9981> Bind needed, attempting Bind for DB OBJSRVRX USING /usr/lpp/frn/bin/BIND/@FRNXLBLB.BSP . <2001-06-24 00:13:39.28 Type: ERRLOG File: lbsydb.c Line: 1633 Func: LbsBindDB Action: LbsBindDB RC: -40 MsgCode: 9861> FRN9861A: Database operation type Bind for DB OBJSRVRX USING /usr/lpp/frn/bin/BIND/@FRNXLBLB.BSP, sqlcode = -40 for the LAN-based object server finished. Return code = -40. . <2001-06-24 00:13:39.38 Type: ERRLOG File: lbscmn.c Line: 9210 Func: LbsInitializeDBManager2 Action: LbsBindDB RC: -40> Bind needed, attempting Bind for DB OBJSRVRX USING /usr/lpp/frn/bin/BIND/FRNXLBYD.BND . <2001-06-24 00:13:39.50 Type: ERRLOG File: lbsydb.c Line: 1633 Func: LbsBindDB Action: LbsBindDB RC: -31 MsgCode: 9861> FRN9861A: Database operation type Bind for DB OBJSRVRX USING /usr/lpp/frn/bin/BIND/FRNXLBYD.BND, sqlcode = -31 for the LAN-based object server finished. Return code = -31. The lbsmonit.h apparently has FRNXLBYD.BND in upper case while the actual file is lower case. It is believed this only fails for upgraded systems. ************************* APAR NUMBER: IR46521 COMPONENT: Object Server APAR DESCRIPTION: When customer select userid properties to reset password. They would highlight the password field, type the new password, then choose ok. The 1st character of the password would get lost.. I ************************* APAR NUMBER: IR44486 COMPONENT: Library Server APAR DESCRIPTION: Customer had a Content Manager V6.1 Library Server "hang". Analysis of the frndiag.log showed the following message: ..... 2000-10-12-12.33.45.066090 Instance:FRN Node:000 PID:26404(frnxllch) Appid:none LIBRARY SERVER LL Probe:1 <2000-10-12 12:33:45.06 Type: ERRLOG File: frnpllch.c Line: 1833 RC: 7474 MsgCode: 7474> FRN7474A: An internal system problem occurred. ..... The FRN7474A is issued by frnpllch at this point because there is an error in receiving an order from the isolator api that we call. This APAR is created because the error code that is returned is not captured in the FRN7474 message. FIX DESCRIPTION: Changed the log statement to log the IsoReceive error message. ************************* APAR NUMBER: IR44489 COMPONENT: System Admin APAR DESCRIPTION: Expired time is locked by the last person/userid who logged onto sys admin... Customer has to logged on sys admin, then log off and log on again to have the rights to update/change expired time.expired time is suppose to be locked by the last person who updated not the last userid/person who logged onto sys admin.side note: expired time, once enabled can't be disabled. FIX DESCRIPTION: When scheudling a util, we were using the logged on userid and not the scheduler userid. ************************* APAR NUMBER: IR44258 / AIX COMPONENT: Library Server APAR DESCRIPTION: Jobber dies from DLL build processing making the library server unavailable. This is due to the size of the .sqc file exceeding a limitation in memory usage. FIX DESCRIPTION: Processing a very complex search on a database canoverflow our 10k buffer search clause. I have modified thecode to check the complexity of the search being processed.If the complexity is above a certain level, we will not try to process that search. ************************* APAR NUMBER: IR45048 COMPONENT: Library Server APAR DESCRIPTION: The Library Server does not issue an error message if the wrong userid is used to start it. If the wrong userid is used to start the Library Server, it initializes and presents the user with the "Accepting Requests" message. However, users get error message " FRN6740A: The logon was not successful because of error 6056 " when they try to log on. The system admin client gets a " FRN8571A: =6056 " message. The frndiag.log shows -204 SQL return codes (undefined name). To determine the correct userid, open the DB2 control center and view the Library Server database tables. The schema name is the userid that should be used to start the Library Sever. The Object Server issues the following error message when started with the wrong userid: FRN9745A: Unable to update database, because it is the incorrect database version, -1. FIX DESCRIPTION: Added a dymamic sql query that will fail under the wrong id ************************* APAR NUMBER: IR46389 / AIX COMPONENT: Library Server PMR NUMBER: 19126,519,000 APAR DESCRIPTION: The frncrtdb.lib.sh shell script calls frncrtd1.sh. If on a normal system, frncrtd1.sh calls frnxinlc.db2 to create the db tables. However, if on a DB2 EEE system, frncrtd1.sh calls frnxinle.db2. The frnxinle.db2 shell script has two problems: 1. The command to create an index on SBTFOREIGNPARTSNMS has a bad column name in the command. This causes a rc=8930, and the db creation process halts. 2. The frnxinle.db2 shell does not seem to be updated. It creates a CM v6.1 database (it shoudlcreate a v7.1 database). The new tables are missing (ie SBTCOPIES) . When the bind runs, "table not found" error messages are received. FIX DESCRIPTION: Modified the aix db create script to add the missing tables and fixed the index column names ************************* APAR NUMBER: IR43486 / AIX COMPONENT: Library Server APAR DESCRIPTION: Warning errors from DB2 during a PREPARE for a dynamic query when retrieving an object resulted in a failure of the retrieve. FIX DESCRIPTION: In the case of receiving a warning code during a PREPARE, log the SQLCODE but attempt to continue the processing (it more often then not will be successful). ************************* APAR NUMBER: IR46025 COMPONENT: CS CAPI Toolkit APAR DESCRIPTION: Custom written user application terminates with a Dr. Watson error message indicating module frnndamn.exe, and text of: 'Exception: Access violation (0xc0000005)'. The application is used to load data into CM. Increasing FRNHEAPSIZE delays the access violation, setting it very small causes it to happen almost immediately. Tracing the application revealed that the access violation is actually in frnnslap.dll (which is a part of frnndamn.exe). This appears to be a memory leak in folder manager. FIX DESCRIPTION: 1. Daemon has been changed to handle error correctly when no more memory is available instead of continuing, and thus causes the crash. 2. Folder MAnager API, SimLibAddFolderItem, and several other API's, has been changed to always free the response block regardless of return code is good or bad. This will avoid memory leak when error occurs. ************************* APAR NUMBER: IR44629 COMPONENT: Library Server APAR DESCRIPTION: OLD PROBLEM DESCRIPTION (NOT ACCURATE): Migrator fails to delete a 'D' status object. The following messages are posted: FRN9763A, FRN9763A,FRN9767A, FRN9763A FRN9763A and FRN9860A rc9763. The row in BASE_OBJECTS shows the volume where the file is located does not belong to the collec- tion. This condition was caused by two quick requests for collection changes without running destager in between the requests. In the case of this customer, collections are assigned by index class. A user reindexed the same itemid twice quick enough that destager did not have a chance to run. The second collection change used the undestaged row from the first collection change and then flagged it with 'D' status which then shielded it from the destager. The problem is that the collection change process copies the source row without altering the file name and path. It is expected that destager will do this. Destager never gets a chance so the row is left with an invalid volume path and 'D' status from the second collection change. Migrator can not mount this (ADSM) volume and fails. The impact is that there are 6 messages per error which creates a large frndiag.log and the through put of migrator can be adversly affected. --------------------------------------------------------------------------- NEW PROBLEM DESCRIPTION: Migrator fails to delete a 'D' status object. The primary message is FRN9860A rc9763 but FRN9763A and FRN9767A usually preceed it. The row in BASE_OBJECTS is identified by the 'object' in the FRN9860A message which is a timestamp that replaced the original OBJ_OBJECTNAME at the time the delete status was changed. The OBJ_VOLUMEID will relate to an ADSM/TSM SMS definition that will not belong to the storage class. This condition is the result of rapid collection changes (at least two) where the original source volume was ADSM/TSM and destager is either disabled or the collection change occurs so rapidly that destager is inbetween cycles. The first collection change created a 'U' status object that would normally be handled by destager. It is the second collection change that processes the 'U' status such that the odd 'D' status BASE_OBJECTS row is created. This APAR number will be used to identify this issue in CM 6.1 and 7.1 fixpacks/CSDs (AndyM). FIX DESCRIPTION: During reindex, if the object's status is 'U', there is no need to delete the file in the original storage system. ************************* APAR NUMBER: IR44519 COMPONENT: System Administration PMR NUMBER: 08852,075,724 APAR DESCRIPTION: Userexit dll names should be restricted to length of 8 characters to avoid confusing errors FIX DESCRIPTION: Modify dialog to not allow more than 8 characters in userexit dll name ************************* APAR NUMBER: IR46455 COMPONENT: CS CAPI Toolkit PMR NUMBER: 48321,090,618 APAR DESCRIPTION: If the SimLibReadAttr API call is issued with the System Attribute OIM_ID_ITEM_CREATE_TIMESTAMP, the call returns an empty string instead of the expected timestamp. FIX DESCRIPTION: If the SimLibReadAttr API call is issued with the System Attribute OIM_ID_ITEM_CREATE_TIMESTAMP, the call returns an empty string instead of the expected timestamp. A type cast was missing during an if statement comparision therefore the comparision always fail. This problem was present in the functions FM_ReadAttr and FM_WriteAttr2. Add a type cast so both data values can be compared using the same data type. **************************************** Fixpack 710.1 **************************************** APAR NUMBER: IR44458 COMPONENT: Toolkit / 5648D0451 APAR DESCRIPTION: When calling simlibsearch with the null (0 Length) search string client receives FRN6772A error message. FIX DESCRIPTION: In LibPackageHostVarValue function (in frnpllrs.c) set string length to 1 when host var's attrType is varchar and host var string is null and the usOperator is relational. ************************* APAR NUMBER: IR44921 COMPONENT: Toolkit / 5648D0451 APAR DESCRIPTION: FRN8635A - Timeout occured waiting for the Library Server to respond to a request. The program times out at the retreive step (ExtRC=8635). DB2 traces show that the update and retrieve are sent to DB2 as two different units of work, so the retrieve ends up waiting for the update to finish, which won't happen until the retrieve finishes. A classic deadlock / timeout situation. FIX DESCRIPTION: Depending whether we are in a transaction or not, derive the proper Library Client session handle to be used within SimLibqueryPartExtSrch. This APAR also fixes a login problem where ProcessACLRows causes an access violation during logon. ************************* APAR NUMBER: IR45091 COMPONENT: Toolkit / 5648D0451 APAR DESCRIPTION: API SimLibCreateItem cannot handle system defined attributes and failed with error FRN6216. FIX DESCRIPTION: Changed SimLibCreateItem code to allow handling of system-defined attributes. ************************* APAR NUMBER: IR45268 COMPONENT: Object Server / 5648D0411 APAR DESCRIPTION: Remote data migration from OS/2, WinN, or AIX v241 to CM 61 or later leaves files on the Target Object Server's "staging" volume without DB entries due to a design change in transaction logic in CM61. These HOMELESS files reside on the "staging" volume and are not destaged/archived to other volumes. The customer subsequently has their "staging" volume fill up. FIX DESCRIPTION: Remote migration cleanup on CM61/71 destages/archives the HOMELESS files to other volumes and the "staging" volume then reclaims space. On the Object Server, the environment variable FRN_DESTAGEALL=1 must be set for the destager to handle these unique files. ************************* APAR NUMBER: IR45227 COMPONENT: Common Services / 5648D0431 APAR DESCRIPTION: When setting environment variable DB2CODEPAGE to a different value on a CM client system and then logging on to the server, the client receives error "FRN8180A The system cannot find either the message file or the message in the message file". FIX DESCRIPTION: CM codepage conversion relies on the DB2 OSS layer. Setting this DB2 environment variable changed the behavior of CM. We fixed it by removing the dependency on DB2CODEPAGE in CM codepage conversion. ************************* APAR NUMBER: IR45234 COMPONENT: Object Server / 5648D0411 APAR DESCRIPTION: Replication within same server had permission problems in staging volume. Subsequently the files were never replicated to the archive volume. FIX DESCRIPTION: Migrator code changed to fix this problem. ************************* APAR NUMBER: IR45320 COMPONENT: Toolkit / 5648D0451 APAR DESCRIPTION: When hObj->sVersion is not 0, any media object cannot be retrieved. SimLibOpenObject gets rc= 8682 FIX DESCRIPTION: In SimLibOpenObject code, after the LibRetrieveItemPartAsync finishes the RETRIEVE_ORDER_CL order creation, if sVersion > 0, then reset the szElementKey field in the element table using sVersion -1 as the internal version number so the szElementKey field will match the media object name sent over from the Object Server instead of unmatching object name in the element table. ************************* APAR NUMBER: IR45627 COMPONENT: Toolkit / 5648D0451 APAR DESCRIPTION: After loading one or more objects into CM, either SimLibUpdatePartExtSrch or SimLibUpdateObject is called to update the content of the objects. If the size of any object is greater than 67000 bytes, SimLibUpdatePartExtSrch or SimLibUpdateObject will fail on AIX with return code = 6256, or SIM_RC_RESIZE_FAILED, and returned hErrLog = 6395, or SIM_RC_WRITEOBJECT_FAILED. If Folder Manager trace is turned on, it will report a return code of 6256 by SimLibWriteObject. As a result, the update activity fails. FIX DESCRIPTION: Function CmnRetrieveFileData called fread() incorrectly. The third parameter of fread() function call should be 1, not the size of the file. Only the second parameter should be the size of the file, but not both the second and third parameters. ************************* APAR NUMBER: IR45951 COMPONENT: Toolkit / 5648D0451 APAR DESCRIPTION: The 16-byte itemID's currently are generated by clients randomly. This may cause performance problems for customers if the number of itemId's is huge. To solve that performance problem, a Library Client user exit is created to allow users to use their own algorithms to generate new itemID's in a more sequential manner. FIX DESCRIPTION: The default userexit is a no-op function and simply returns 0 to the caller. A sample userexit (commented out) is provided in this userexit. To use the sample method, users can simply un-comment it and modify it anyway desirable by the users, and build the userexit again. So the default behavior of this userexit is business as usual, that means no impact to existing customers, unless the userexit is modified and rebuilt by the customers. User exit: LibCreateItemID (generate item ID user exit) Format: LibCreateItemId (pszItemID,pszWorkStationID,sIndexClassID,pszUserID, sSemType) PURPOSE: When an application creates an item, Content Manager assigns the item several system-defined attributes. One of those system-defined attributes is an item identifier (item ID), which is randomly created by the Library Server Client. This user exit allows you to change the way the Library Server client generates item IDs. You can use the sample in this user exit to have the library client create sequential item IDs. To use the sample, remove the comments tags in the code and recompile. You can also modify the sample code and add your own algorithms. The library client encrypts the password that is associated with the user before it is transmitted to the library server. Use LibEncryptPassword if you want to provide your own encryption algorithm or eliminate the password encryption. Note: If you don't "uncomment" the sample code, the CM library server client will create item IDs as it is originally designed to do; random ID generation. PARAMETERS: pszItemID: CHAR - input and output A parameter passed in by the internal code. pszWorkStationID: CHAR - input The character string of the host name. The maximum length for this string is 18 bytes, which includes the null terminator. sIndexClassID: SHORT - input The index class identifier. pszUserID: CHAR - input The NULL-terminated character string that specifies the password for the user ID. This parameter is not case sensitive. sSemType: SHORT - input The semantic type of the item to store. The semantic type of an item indicates its general classification. Use a 2-byte integer. The predefined semantic types are: SEMTYPE_ACL Indicates that the item is an access list. SEMTYPE_ANNOTATION Indicates that the item is an annotation. SEMTYPE_DOCUMENT Indicates that the item is a document. SEMTYPE_FOLDER Indicates that the item is a folder. SEMTYPE_HISTORY Indicates that the item contains history events. SEMTYPE_MGDS Indicates that the item contains MGDS data. SEMTYPE_NOTE Indicates that the item is a note. SEMTYPE_USER Indicates that the item is a user item. SEMTYPE_WORKBASKET Indicates that the item is a workbasket. SEMTYPE_WORKFLOW Indicates that the item is a workflow. SEMTYPE_NONE Indicates that the item has no semantic type. All other values are for use by applications. The library server does not interpret them, so they can be used by the application in any way you want to. RETURN VALUES: The function returns a value of SHORT with zero for SUCCESS and -1 for FAILURE. COMMENTS: The function is located in the FRNPLCPW.C module. The LibCreateItemID function is a single function in FRNWLCPW.DLL. You can replace FRNWLCPW.DLL with your own FRNWLCPW.DLL. ************************* APAR NUMBER: IR46219 COMPONENT: Object Server / 5648D0411 APAR DESCRIPTION: Object Server Recovery deletes files not found in the database. Normally this is recommended, but in the case of APARS IR45028 and IR45268, these files will unintentionally be left in an abnormal state and are not in the database. So deleting these files before IR45268 is applied and run will permanently remove these files. FIX DESCRIPTION: The default behavior of Object Server recovery NO LONGER deletes files not found in the database. If you wish this behavior to happen, then you must set the Environment variable FRNRECDELETE=1 and run recovery to delete the files. *********************************************** 4. Documentation APAR Fixes with Descriptions *********************************************** APAR NUMBER: IR46244 COMPONENT: Admin APAR DESCRIPTION: CM71 online help information does not reflect (contrary to CM61 and earlier releases) that the ctrl key must be depressed in order to select/add an user to multiple user groups from the listed groups. FIX DESCRIPTION: In the Content Manager for Multiplatforms System Administration Version 7 Release 1, the following online help should be as follows: New User panel: Select the user groups to associate with the user. To select multiple user groups, hold down the Ctrl key when selecting the groups. ************************* APAR NUMBER: IR45737 COMPONENT: Admin APAR DESCRIPTION: Customer reports that the System Admin Client does not show correct data on the Staging area as system is running. They have to shut down and reboot to show correct data. The SysAdmin Client is working as designed in this case, and the customer needs to use the command interface (frncmd) to obtain this information. FIX DESCRIPTION: In the Content Manager for Multiplatforms System Administration Guide Version 7 Release 1, the following information is added: Chapter 6, Server Control, "Monitoring and controlling the servers" For AIX, or Windows NT: To monitor the current status of the staging area, use the command utility with the status command (see Command utility on page 44). For example, to monitor current status on the object server named LDVOSAIX, enter: frncmd ldvosaix status ************************* APAR NUMBER: IR46475 COMPONENT: Admin APAR DESCRIPTION: In the CM v7.1 SysAdmin Guide, Ch 11, p118, the imlclpix command is described as being used to "Remove all index entries from an existing index on the text search server." This statement is somewhat half true ... it should say "This is a cleanup function for library server consistency. This removes all index entries from an existing index on the text search server which no longer exists in a CM library server" Command line, imlclpix -h, gives the correct information. FIX DESCRIPTION: In the Content Manager for Multiplatforms System Administration Guide Version 7 Release 1, the following information is corrected: page 118, Chapter 11: Text search administration The text under the syntax diagram for "Delete Content Manager documents from an index (icmlclpix)" should read: Purpose: This is a cleanup function for library server consistency. This removes all index entries from an existing index on the text search server which no longer exists in a Content Manager library server. ************************* APAR NUMBER: IR46263 COMPONENT: Object Server APAR DESCRIPTION: The Planning and Installing Content Manager Version 7.1 manual (GC27-0864-01) discusses the need for an entry in the FRNOLINT.TBL for the client messenger under certain circumstances. (See Chapter 9, 'Configuring network communication for Content Manager components', topic 'Understanding Network Tables', sub-topic 'Client toolkit notes', page 182.) However, the correct syntax and parameter definitions for the FRNODAOS statement, and when you need to use them, are not in the in the 'Network table file syntax' section of chapter 9, which also does not document FRNODAOS as a reserved word. FIX DESCRIPTION: In the Content Manager for Multiplatforms Planning and Installing Guide Version 7 Release 1, the following information is corrected and added: Page 182, Client toolkit notes: Delete the following text: Optionally, specify the port and IP address for the client messenger to use for object storage and retrieval. Specifying the IP address is useful when the client contains multiple communications cards or when there is a router that performs NAT (network address translation) between the client and servers. In the latter case, the external address must be the IP address. And replace with the following text: In some configurations it may be desirable to place the CM toolkit and its applications on one side of a firewall while the servers are the other side of a firewall. Also, it might be desirable to choose which communications network card is to be used for object transfers to and from the object server, such that object transfer occurs on a specified network, given that the client workstation has access to multiple networks. Accommodating both such configurations is done through the creation of a local client network table entry on the client toolkit workstation. Use the network table generator to ADD LOCAL CLIENT specify Port specify IP Address In the case of a workstation attached to multiple networks, specify the IP address of the communications card that is desired for object transfer to and from the object server. Port, in this case, can be any port in the valid port range specified in In the case of a firewall, it can be somewhat restrictive. The firewall must use static NAT, and not dynamic. With static NAT, each client must have an internal ip address, known to the servers, mapped to an external ip address. With such a mapping, use the network table generator at the client toolkit workstation to add the local client definition. Specify the internal IP address. The port to be specified is usually the one in which the firewall is configured to permit. Most administrators find that using the same port number as the library server works best, given that the client machine doesn't contain a library server itself. That way, the firewall can be configured to allow traffic for that port number and it will work for both the traffic going to the library server and from the object server. Finally, in both the firewall and multi-honed configurations, on a Windows toolkit platform, the administrator must determine whether or not there will be multiple applications run on the workstation. For example, Does there exists the CM client application and a custom application, or sys admin on the workstation? This situation can be called a multi-process configuration, where multiple applications share the client configuration. Conversely, if only one application is to be run on the workstation, it is a single-process configuration. It is important to determine which environment the workstation will be running and set environment variables as follows: * single-process configuration - add FRNSINGLEPROCESSCLIENT to system environment variables setting its value to 1. * multi-process configuration - add FRNSINGLEPROCESSCLIENT to system environment variables setting its value to 0. Additionally, update the system environment variable FRNHEAPSIZE=2048. In this environment, the applications will be sharing the client messenger daemon and requests/responses to/from the library server and in some cases, objects to/from the object server will be placed in a shared memory heap. The setting of a larger shared memory heap is required. 2048 translated to 4K X 2048 = 8M of shared memory. This should be sufficient for most requests, responses, and objects. But if the application starts receiving out of memory failures, (because of large objects or larger results sets from queries), the FRNHEAPSIZE should be increased. Configuring for the default client messenger Nothing further needs to be configured in the default toolkit configuration. Page 195, add the following text to Network table file reserved words: FRNODAOS Page 195, add to definition column which states For library servers and object servers, the server name is the same as the database name. For the client messenger, the name is FRNODAOS ************************* APAR NUMBER: IR42892 COMPONENT: Object Server APAR DESCRIPTION: The message FRN9020A is used to reflect an error received from the protocol that is in use. This can TCP/IP as well as APPC. The message text needs to be changed to reflect this usage. FIX DESCRIPTION: In the Content Manager for Multiplatforms Version 7 Release 1 Messages and Codes manual, the following message is updated with TCP/IP information: FRN9020A Communications Manager/2 or SNA Server/6000 returned an advanced program-to-program communications (APPC) error or transmission control protocol/internet protocal (TPC/IP) error. Explanation: Communications Manager/2 or SNA Server/6000 issued an APPC or TPC/IP error, which caused the client or server program to stop processing. Component: Communications Isolator #Define: ISOERR_APPC_ERROR Action: Check the error logging facility to see which APPC or TCP/IP error code was issued. See the error codes in the Communications Manager/2 APPC publication or TCP/IP publication and correct the problem. ************************* APAR NUMBER: IR44174 COMPONENT: Object Server APAR DESCRIPTION: The Planning and Installation Guide, GC26-9831-00, Chapter 4, Installing Content Manager components on AIX, Step 4.Installing the object server on AIX, states the frnbind.obj should be run in #7. under the Oracle set of instructions. Development states this is not true FIX DESCRIPTION: In the Content Manager for Multiplatforms Planning and Installation Guide Version 7 Release 1 Pages 62-65, Step 4. Installing the object server on AIX Page 65, under Table 40, delete the following item (because you do not need to bind the database for Oracle): 7. If you did not create a database while using the wizard, bind your existing database. Enter: frnbind.obj ************************* APAR NUMBER: IR44499 COMPONENT: Object Server APAR DESCRIPTION: OBJECT SERVER WITH A LARGE STAGING DIRECTORY BUT NOT ON A DEDICATED SPACE WILL ENCOUNTER EXTENDED STARTUP AND RECOVERY If a large staging area is to be used, it is recommended to utilize a dedicated staging area. If the area is not a dedicated partition, the object server recovery and startup, which performs an automatic recovery, will be extended due to the large number of files that will be reviewed. Installations migrating from previous versions of Visualinfo should strongly consider implementing this option. In addition to setting STA_DEDICATED='Y' in the BASE_STAGING table, modify the value of STA_PATH to the newly defined dedicated partition. FIX DESCRIPTION: In the Content Manager for Multiplatforms Planning and Installation Guide Version 7 Release 1, the following information is added: Page 52, Step 4 Installing the Object Server on Windows, Fast object server recovery is now Fast object server startup Page 65, Step 8 Installing the Object Server on AIX, add the following information: Fast object server startup The object server should be started and shutdown once after installation. If the staging area is a dedicated space, you see that a column added to the object server table BASE_STAGING called STA_DEDICATED. This column is "dedicated" because it is only used by the object server, and only staged blobs are stored in it. To update the server: 1. Log on to object server admin user ID Example: su - osadmin 2. Connect to the object server database Example: db2 connect to objsrvrx 3. Update the BASE_STAGING database table, column STA_DEDICATED, using the following query: Example: db2 UPDATE BASE_STAGING SET STA_DEDICATED = 'Y' This command queries the file system to find out how much space is being used on the file system. This avoid a time consuming filter of the files in staging. The result is a close to constant startup time independent of the number of files in the staging area. If STA_DEDICATED is set to 'N' then the startup time depends on how many files are in the staging area. Never set STA_DEDICATE to 'Y' if the file system or partition holds files other than object server blobs. In the system administration program, do not allocate the staging space to be the size of the entire partition. Allow for some level of error as file systems report of space is not always completely usable due to overhead in the file systems. For most file systems the error is less than 10%. A good starting point for a 10,000 MB partition would be to set staging space to 9,000 MB with a purge start of 90% and a stop of 80%. These are only guidelines -- you should monitor the staging area as it approaches its full state and during periods of intense activity be able to ensure the staging file system does not get completely full at any time. If the file system is never close to being full, then the values can be increased. Destaging and purging problems might result if the values are set incorrectly. ************************* APAR NUMBER: IR46023 COMPONENT: Object Server APAR DESCRIPTION: Customer reports that Object Server start up time is 5.5 hours. The Object Server appears to be looking at every file in the staging area ( ubosstg ) which has approximately 190,000 files. FIX DESCRIPTION: In the Content Manager for Multiplatforms Planning and Installation Guide Version 7 Release 1 Page 65, Step 8 Installing the Object Server on AIX, add the following information: Fast object server startup The object server should be started and shutdown once after installation. If the staging area is a dedicated space, you see that a column added to the object server table BASE_STAGING called STA_DEDICATED. This column is "dedicated" because it is only used by the object server, and only staged blobs are stored in it. To update the server: 1. Log on to object server admin user ID Example: su - osadmin 2. Connect to the object server database Example: db2 connect to objsrvrx 3. Update the BASE_STAGING database table, column STA_DEDICATED, using the following query: Example: db2 UPDATE BASE_STAGING SET STA_DEDICATED = 'Y' This command queries the file system to find out how much space is being used on the file system. This avoid a time consuming filter of the files in staging. The result is a close to constant startup time independent of the number of files in the staging area. If STA_DEDICATED is set to ¡¦N¡¦ then the startup time depends on how many files are in the staging area. Never set STA_DEDICATE to yes if the file system or partition holds files other than object server blobs. In the system administration program, do not allocate the staging space to be the size of the entire partition. Allow for some level of error as file systems report of space is not always completely usable due to overhead in the file systems. For most file systems the error is less than 10%. A good starting point for a 10,000 MB partition would be to set staging space to 9,000 MB with a purge start of 90% and a stop of 80%. These are only guidelines -- you should monitor the staging area as it approaches its full state and during periods of intense activity be able to ensure the staging file system does not get completely full at any time. If the file system is never close to being full, then the values can be increased. Destaging and purging problems might result if the values are set incorrectly. ************************* APAR NUMBER: IR38764 COMPONENT: Library Server APAR DESCRIPTION: VisualInfo - Digital Library - trace - FRNTRC - FRNTRACE The discussion contained in Messages and Codes Chapter 3 Tracing Facility does not clearly identify the relationship of the controls (and variables) to the effect on the trace. For example, the variable FRNTRACELEV and those that follow almost seem to be unrelated to the FRNTRACE function by the way it is worded. The file 'frntrace.ctl' does not exist on NT. How does FRNLOGMAXMB relate to the -l parm for FRNTRC, and if there is no connection, what does FRNLOGMAXMB really do? OS2 is referred to repeatedly in the discussion for FRNTRACE but it is not functional on the OS2 platform for VI. FIX DESCRIPTION: In the DB2 Digital Library Messages and Codes Version 2 Release 4 manual, the following information is deleted: Page 6, Chapter 2. Tracing facility Deleted text: Tracing on AIX and Windows NT: Digital Library also provides additional features through its Independent Trace Facility (ITF) for collecting trace information. Pages 6 through 19: all information describing the following options is not valid and is deleted: * [-m mask] mask= products. event-types. components. functions * [-e maxSysErrors] Stop tracing after maxSysErrors have been recorded * [-r maxRecordSize] Truncate trace records at maxRecordSizebytes * [-rc return-code] Suspend tracing when return-code is encountered * [-p processId[. threadId]] Process ID and thread ID to trace. Thread ID does not apply to AIX. * [-c companion-id] The companion ID to trace in a server environment * [-s | -n | -mmio | -f fileName] Send trace records to shared memory, native trace facility, performance monitoring hardware, or a file * [-crash crash-mask] crash-mask = product-id. event-types. components. functions. trace-point Cause the program to exit when a specific trace point is reached ************************* APAR NUMBER: IR44974 COMPONENT: Library Server APAR DESCRIPTION: There is a need to document the limits of what the Optimization utility can be expected to do for the customer. For medium to large installations to get the best performance from the VI databases the customers' DBA will need to manage the database through the database utility functions that come with the database product FIX DESCRIPTION: In the Content Manager for Multiplatforms System Administration Guide Version 7 Release 1, the following information is added: Page 72, Chapter 9, Database installation and maintenance The optimization utility The optimization utility that is provided with VisualInfo is intended to be used for a small to a moderately sized database. This utility performs the basic optimization functions necessary to keep your server database performing optimally. As your system grows and becomes more complex, it is recommended that optimization of the server database be performed by a qualified database administrator using the tools provided by your database product. As a rule of thumb you may use the following guidelines to determine when your system has gotten to complex for the optimization tool of VisualInfo: System with more than 100K items System with more than 20 index classes System containing multiple database indexes on an index class ************************* APAR NUMBER: IR42533 COMPONENT: Library Server APAR DESCRIPTION: For the RELCRITERIASTRUCT used by Ip2SuspendItem, the doc says that itemDeadlineWB is ignored if the tsExpDateTime is NULL. TL says this isn't true and that an itemid needs to be filled in anyway (a copy of what's in itemidReadyWB should do). The doc should be modified to reflect this. FIX DESCRIPTION: In the Content Manager for Multiplatforms C Application Programming Reference for Workstation Vol 1 Version 7 Release 1 Page 930, RELCRITERIASTRUCT (Release Criteria Data Structure) RELCRITERIASTRUCT description of itemidDeadlineWB has changed to: ITEMID - input The item ID of the workbasket to send the suspended item to if the expiration time criteria are satisfied. Add the following text: This field cannot be NULL. If you want to release the suspended item due to other criteria (e.g., not deadline), then you need to copy the item ID from the itemidReadyWB or pitemidReadyWBArray to itemidDeadlineWB parameter. Delete the following text: This field is required if the usReleaseType field contains the value OIM_SUSP_ANY_CLASS or OIM_SUSP_ALL_CLASS, and if pitemidReadyWBArray has not been specified. Otherwise, this field is ignored. ************************* APAR NUMBER: IR45849 COMPONENT: Library Server APAR DESCRIPTION: Documentation in the Application Programming Guide for Windows v7.1 (APG) is incorrect concerning proper interpretation of the correct column width returned by LibGetQueryReplyInfo. with more detail of: Use the function DmnFieldWidth (in frnndapi.lib) to intrepret the correct column width regardless of data type. FIX DESCRIPTION: In the Content Manager for Multiplatforms Application Programming Guide Version 7 Release 1, the following information is added: Use the function DmnFieldWidth (in frnndapi.lib) to intrepret the correct column width regardless of data type. For example: if (pCol->col_type %2==1) { if (pCol->col_width<=1) return; /* Skip over the NULL byte indicator */ pbRowData++; usFieldLen = DmnFieldWidth(pCol->col_type,pCol->col_width,pCol ->col_desc_len, bFormatted) - 1; } else usFieldLen = DmnFieldWidth(pCol->col_type,pCol->col_width,pCol ->col_desc_len, bFormatted); ************************* APAR NUMBER: IR45861 COMPONENT: Library Server APAR DESCRIPTION: Content Manager v7.1 Application Programming Guide for Windows Appendix B - User Exit Routines The customer is building a SaveRecordUserExit module, and is having trouble getting the USEREXITSTRUCT to map properly to the actual data. FIX DESCRIPTION: In the Content Manager for Multiplatforms Application Programming Guide for Windows Version 7 Release 1 Appendix B, User Exit Routines, Client for Windows User Exits itemidWorkflowid field has changed from USHORT to ITEMID ************************* APAR NUMBER: IR46064 COMPONENT: Library Server APAR DESCRIPTION: Chapter 12 of 'Planning and Installing Content Manager' provides instructions for using the utility for migrating the library server database to the CM 7.1 version. The instructions state to run frnnmiut.db2.exe from the CD. This does not work. The executable and other files must be copied from the CD to the hard drive to be executed. A proceedure naming what files must be moved is required and the softcopy version of the pubs should be updated. FIX DESCRIPTION: In the Content Manager for Multiplatforms Planning and Installation Guide Version 7 Release 1 Chapter 12. Migrating a Windows or an AIX Content Manager database page 237, Migrating a DB2 library server database on Windows, the following information is added to Step 2: Copy the migration utility (frnnmiut.db2.exe) and the bind file (frnomiut.bnd) into a local directory on the Library Server to be migrated. These two files must reside in the same directory, and the user must have read/write privileges in that directory. Run the migration utility from your local drive. page 242, Migrating a DB2 library server database on AIX, the following information is added to Step 2: On AIX,'su' to the LSADMIN user ID. Create a temporary directory where LSADMIN has read / write privileges. Copy all the files from /AIX/migrate/db2 and /AIX/migrate/db2/bind from the installation CD to the temporary directory (all the files must be the same temporary directory). Run frnxmiut.db2 from the temporary directory. ************************* APAR NUMBER: IR42479 COMPONENT: Library Server APAR DESCRIPTION: For both the C version (COMPID 569725044) and the Java Sysadmin, if the library server configuration is set for a single logon per user, the utilities that can be scheduled such (as expired time) will post FRN7213A. The utility requires the ability to make a second connection when the utility starts. The chapter in System Administration Guide ' System setup and system maintenance' fails to mention that library server configuration must be set for multiple sessions if you want to use the system admin. GUI server utilities. FIX DESCRIPTION: Ensure the library server configuration is not set for 'single session' in the 'Active sessions(Per UserID)' section on the 'Access' page of the library server configuration. ************************* APAR NUMBER: IR47209 COMPONENT: Library Server APAR DESCRIPTION: The customer countered he is running CM 7.1 and that the System Administration Guide says: "Restriction: If your instance of the system administration client is running through the system administration gateway, then the library server and system administration gateway must be logged on at the scheduled time for the utility to run." This is clearly a different statement and implies if the gateway is not used, then you do not have to be logged on. The customer feels the way the product works has changed and if the gateway is not being used, the user does not have to be logged on. FIX DESCRIPTION: In the Content Manager for Multiplatforms System Administration Guide V7.1, the following information is updated: In order for any of the database utilites to run at a scheduled time, the user that scheduled the utility must be logged on to the library server by way of the system administration program. ************************* APAR NUMBER: IR46134 COMPONENT: Library Server APAR DESCRIPTION: Static generation of an expression that includes the NOTIN operator successfully generates a dll, but does not include the sql with the NOTIN expression. Each subsequent submission of the same expression will generate a new static dll, each time excluding the expression. No errors are reported in the frndiag.log FIX DESCRIPTION: In the the Content Manager for Multiplatforms Application Programming Guide and Application Programming Reference, the following information is added: Because of the complex nature of the query searches involving IN or NOT IN, search conditions will not be optimized to static search dlls. ************************* APAR NUMBER: IR47572 COMPONENT: OBJECT SERVER APAR DESCRIPTION: The Object Server Recovery Utility ( frnrecdb.obj ) ignores all command line parameters and uses the parameters hard coded in the script. It uses current environment variables to select the database to run against, and the configuration file to use. FIX DESCRIPTION: In the Content Manager for Multiplatforms System Administration Guide V7.1, the following information is updated: On Page 78 The following statement is incorrect: frnrecdb.obj /d ldvosaix /c ibmconfig /s Change to: frnxlbrc /d ldvosaix /c ibmconfig /s ********************************************* END OF README *********************************************