CallPath Switch Server/2 for Lucent G3

This page was last modified: 2 November 1999, 11:53am

Required Version 2.02
Please note that this service can only be applied to the required version.
To obtain service for previous versions, click here)
Service File g3s1707.exe
Service Readme g3sread.txt has a more detailed description of the defects for this cumulative service, along with instructions on how to apply the service.
Service Level 1707
Size 821957 (802Kb)
Ethernet Service File g3es1707.exe
Ethernet Size 828377 (808Kb)

INTERNAL 4870
Description SS/2 startup.cmd file corrupted

Symptom: If SS/2 is installed on a machine which has had a CallPath Server installed, the startup.cmd file will overwrite the CallPath startup lines and put multiple SS/2 startup lines in the startup.cmd file

INTERNAL 5131
Description New SS/2 backup procedure

Symptom: The SS/2 backup procedure did not successfully back up the SS/2 system.

INTERNAL 5706
Description libpath was not updated for SS/2

Symptom: Certain libraries could not be found when executing SS/2.

INTERNAL 5708
Description galaxy dll's missing.

Symptom: Attempts to configure a switch failed.

Internal 8000
Description Incorrect existing party list in RECONNECT event by calling party

Internal 8000 When the calling party generates a RECONNECT event, the calling party appears in the existing party list as well as the connecting party structure in a two party call. A minor logic error was found in the connect event processing module and has been corrected.

APAR IC24604
Description G3 STLQACD returns all zeros in counts

Internal 7982 Issuing an STLQACD request for types QUERY_PILOT_QUEUED_CALLS or QUERY_PILOT_AGENTS_SUMMARY will always return zero counts. The problem was introduced by recent changes in the G3 raw message decoder and has been fixed.

APAR IC24433
Description G3 Ethernet SWITCH.EXE reports false switch up status

Internal 7923 Starting the G3 Ethernet SWITCH.EXE generates a switch up, switch down, switch up status. The first switch up status was incorrect and the system has been changed to remove it.

Internal 7913
Description Redirect events from bridged alerting parties not reported

Internal 7913 When a call was redirected from the principle station of a bridge group, the redirect events for the bridged parties were discarded. The problem was found in the duplicate event filter used when reading data from the switch link and has been corrected.

Internal 7910
Description Incorrect queue cleanup route reason in routed event

Internal 7910 When a VDN is configured to queue a call to only one skill and later redirects the call after a timeout, an incorrect queue cleanup routed event was reported. The route reason has been changed to use a forward route in this sequence to more accurately reflect the architecture.

APAR IC24432
Description Calling a 16 party bridge causes ERR_REQUEST_COMMBFR (6003) error

Internal 7903 The G3 SWITCH.EXE would go down after receiving a 6003 ERR_REQUEST_COMMBFR error when a 16 party bridge group is called. The cause was found to be a protocol violation in the connected party id field from the 16th alerting event and had a value of 0. Additional defensive code was added to catch this switch fault and handle it using a value of 0x80 instead.

Internal 7881
Description Generate negative response for STLXTND or STLRED to self.

Internal 7881 An invalid directory number negative response is now generated for STLXTND cid_a DN_A and STLRED cid_a DN_A requests where cid_a represents the same party as DN_A. This change was made to be consistent with what happens when an STLMAKE DN_A DN_A request is issued. The STLRED request will only be NAK'ed when redirecting an alerting party. When the STLRED is used to reply to a REQUEST_INSTRUCTION, redirecting back to the originating VDN is allowed, however, a new ROUTED event is generated on the positive response. The new ROUTED event contains the same party information in both the route-from and routed-to fields, but the connection id's are different to show replacement of the existing one with the one created by the STLRED request.

APAR IC24180
Description SWITCH.EXE traps when a bridged call follows coverage to a VDN.

Internal 7868 The G3 SWITCH.EXE will trap/core dump when a call alerts the principle station of a bridged group and redirects to an unmonitored VDN and then to a monitored VDN which forward routes to a second monitored VDN. The Lucent G3 event sequence that must occur is ALERT (principle), ALERTs (bridges), REDIRECT (principle), CALL OFFERED TO DOMAIN (1st VDN), REDIRECT (from 1st VDN), CALL OFFERED TO DOMAIN (2nd VDN). The trap has been fixed during the processing of the 2nd CALL OFFERED TO DOMAIN and the bridged parties left in the call after the call redirects from the principle station are now disconnected.

APAR IC23915
Description Updated revertive call processing to detect fax/modems

Internal 7578 Revertive (predictive) calls that reach a fax/modem are now handled as if they reached an answering machine. Calls that reach a fax/modem will clear with the disconnect reason STL_DISCONNECT_DEST_DETECT versus STL_DISCONNECT_NOT_SPECIFIED as they had done previously.

APAR IC23913
Description Corrupted data in Feature_Invoked events for Agent Login/Logouts

Internal 7530,7573 Recent changes in some of the G3V6.3.1.230.6+ switch software rearranged some of the data fields in the agent login and logout events. The messages were correct according to the ASAI protocol documentation, but the G3 SWITCH.EXE would return the agent ID with a corrupted last character or sometimes return the physical extension as the agent ID. The ASAI decoder component of the SWITCH.EXE has been updated to handle the new messaging formats.

APAR IC23413
Description G3 SWITCH.EXE traps when a VDN service observe is active

Internal 7279 The OS/2 G3 SWITCH.EXE using an ISDN connection to the switch traps when a VDN service observe is active and the link periodically becomes congested. The cause was not handling the routed event for an inbound call that reused a dead call created by the congestion. The problem was isolated to the processing of the CALL OFFERED TO DOMAIN (Routed) event and has been corrected.

Internal 7278
Description Revertive STLMAKE requests directly from ACD splits create dead calls

Internal 7278,T4376 Revertive (predictive) STLMAKE requests originating directly from an ACD (hunt group) without using an VDN will create dead calls. This only happens when the ACD is configured with queueing.

Internal 7263
Description Remove logging of negative responses to disconnect requests

Internal 7263 Removed disconnect negative response logging when the target party hangs up before the request reaches the switch. The messages needlessly fill up the error logs and are not needed anymore.

APAR IC23412
Description STLQPS for feature or indicator returns ERR_RT_RSRC_UNAVAIL (5008) error

Internal 7207 The G3 SWITCH.EXE would go into the overload state after 5000 or more STLQPS requests for features or indicators. The problem was caused by a request tag leak and has been fixed.

APAR IC23411
Description STLMAKE/STLXTND out ISDN trunks to fast-connecting destination produces wrong cid's

Internal 7196 An STLMAKE or STLXTND request using an ISDN trunk to a party that fast-connects (i.e. does not produce an alerting event before connecting) produced wrong connection id's in the connect event. Changes were made to handle this new flow.

APAR IC22699
Description G3 SWITCH.EXE goes down after ERR_REQUEST_COMMBFR (6003) error

Internal 6860 The G3 SWITCH.EXE would go down after receiving a 6003 ERR_REQUEST_COMMBFR error. The cause of the problem was a communication buffer resource leak in the code that does queue cleanup and has been fixed.

APAR IC22698
Description Unexpected CSM errors when conferencing/transfering to external parties.

Internals 6842,6822,6806,6764,6754 The Lucent G3V6 switch software now allows local stations to defined without hardware with coverage paths off-switch. Calling such stations now alert/connect/conference/transfer with ##### instead of their actual extension. This impacted several event transition routines which required changes to handle the new information in the switch messages. Symptoms where unexpected CSM errors with ALT, CC, ROU, QUE tags or newly introduced connection ids in conference or transfer events. Problem is related to APAR IC22240.

APAR IC22697
Description Handle New Lucent G3V6 Cause Codes for the CONNECT event.

Internal 6780 The Lucent G3V6 switch software introduced three new cause codes for the CONNECT event. This only applies to switches running with Definity Release G3V6x.03.01.234.x or higher. The new values are 3/7 (service observing), 3/8 (listen only), and 3/9 (listen/talk). Service observe and listen only causes set the Callpath connect reason to STL_ADD_PARTY_RCV_ONLY and the listen/talk cause to STL_ADD_PARTY_SEND_AND_RCV.

APAR IC22696
Description Type flags not reset when a request tag is reused to monitor a pilot number

Internal 6779 If a switch request tag was used as a temporary monitor for an alerting station and then reused to monitor a pilot number, unexpected CSM errors might occur. Problem was due to some internal flags not being cleared and has been corrected.

APAR IC22241
Description Missing DISCONNECT event for non-monitored or external parties after call ends.

Internal 6721 When call tracking is enabled and the last monitored party for a call disconnects, the DISCONNECT event for the remaining non-monitored or external parties will not flow. This is due to the switch returning a negative response to a take control request. The problem can be recognized by looking at the real-time trace. The trace would show the negative response(NAK) without a corresponding DISCONNECT event.

APAR IC22240
Description Wrong connection id in CONFERENCED event after a NETWORK REACHED event

Internal 6720 The problem scenario occurs when an external party(X1) calls and connects to a monitored party(A). Party(A) then manually conferences another external party(X2) over a non-ISDN trunk. The connection id for party(X2) generated by the NETWORK REACHED event was incorrectly replaced with a new one in the subsequent CONFERENCED event. This problem has been fixed so party(X2)'s original connection id is used and updated with the DNIS from the switch if present.

APAR IC22160
Description Discarded ALERTING event after REDIRECT event.

Internal 6683 The ALERTING event was discarded during the event sequence REDIRECT, CALL OFFERED TO DOMAIN, ALERTING. The condition can be identified by an the error ERR_UNEXPECTED_CSM_STATE with an ALT: in the message. This typically happens when a VDN contains a route-to step to another VDN and generates the REDIRECT, CALL OFFERED TO DOMAIN and the target VDN contains a converse vector step and generates an ALERTING event. The problem was corrected by clearing a conditional when processing the CALL OFFERED TO DOMAIN event after it was set when processing the REDIRECT event.

APAR IC22159
Description PICKED vs. CONNECT event sent when trunk to trunk call CONNECTS.

Internal 6683 When external T1 calling party connects to another external T1 party after a transfer, a PICKED event was sent vs. a CONNECT event. This caused the unpark (pick) call test to be true because the calling and called parties were the same. The test was fixed to disallow this for external parties which are represented by #####'s on T1 trunks and now flows a CONNECT event.

APAR IX83160
Description Intermittent core dump when the G3 closes the Ethernet link.

Internal 6675 When the Definity LAN Gateway closes the Ethernet link to the Callpath server due to a network error or internal switch problem, a core dump would sometimes be generated by the RequestControl process. The core dump was occuring because another process would forcefully remove the request queue from the RequestControl process causing a memory access violation. The problem was found in the cleanup code that runs during shutdown and has been corrected.

APAR IC22158
Description CPSS2 for G3 Ethernet ignores the NOSTART option.

Internal 6675 When CPSS2 would start the G3 switch dependent code, it would immediately exit. The problem was the SWITCH.EXE process would unconditionally try to connect to the Definity LAN Gateway and exit if the link failed to start. The startup sequence was changed to open the link only upon request. This allows the user to configure the host links to SS/2 without the link to the G3 being active or even correctly configured and allows the link to stopped and restarted by request.

APAR IC22157
Description Get ERR_CREATE_CSM when starting G3 switch on WINNT.

Internal 6675 The switch dependent code would generate an ERR_CREATE_CSM error (unable to start call state model) which could only be cleared by a system reboot. The problem was found in the cleanup code after a link failure and has been corrected.

APAR IC21786
Description Get unexpected CSM errors with queued calls routed off-switch by VDN.

Internal 6621 Fixed processing of the CUT-THROUGH/TRUNK SEIZED event to do queue cleanup using ROUTED messages before sending up a NETWORK REACHED message. This change will handle VDN with vectors that queue calls to one or more splits and contain route-to steps that route the call to another switch for load balancing.

APAR IC21793
Description ISDN or Ethernet Link to G3 PBX fails to start on correctly configured server

Internal 6617 Added error code ERR_INVALID_CRV_LENGTH to catch situations where the Callpath Server configuration is verified to be correct, but the link will not start. This can happen when the ASAI station on the PBX is not configured correctly. The text of the new error code contains a description of the correct ASAI station settings which customers can quickly check and fix on their own before involving IBM or Lucent support.

APAR IC21785
Description No response to STLTRAN request after STLALTR completes.

Internal 6608 Changed the transfer request processing (STLTRAN) to disallow transfers from a call with only one party in it. This can happen when A calls B, B answers, B extends (STLXTND) to C which is busy, STLALTR B-A B-C to reverse the held and connect states of the calls at B, and then transfer (STLTRAN) the second call to the first one. This flow exposed a Lucent switch bug where no response was sent back from the transfer request and potentially causing a hang at a client application.

APAR IX81036
Description PARTY_STATUS_MSG for agent status query always returns AGENT_LOGON (0).

Internal 6594 Fixed the STLQPS using the STL_PARTY_FEATURE_STATUS for AGENT_ACTIVITY to correctly send back the agent state information. The problem was unique to AIX and was traced down to a byte-swapping issue during the message encoding process. The symptoms were zeros returned in the party status messag fields vs. the correct values. This error also occurred for STL_FORWARD and STL_PARTY_INDICATOR_STATUS queries.

APAR IX81781
Description Support Service Observe CONNECT events.

Internal 6573 Changed processing of the CONNECT event to identify service observe connects by setting the connect reason to STL_ADD_PARTY_RCV_ONLY. This update completes the 5829 change listed below. The 4912 change was removed because the flow it was designed to fix was a VDN service observe connect. Limitations: The event stream from the switch for service observe connects is identical to some call picked flows. Full support where the connect reason is 100% reliable will require a new cause code from the Lucent switch which is not yet available.

APAR IC21746
Description Get unexpected CSM errors while party is on HOLD and/or RECONNECTS.

Internal 6557 Fixed processing of the ALERTING, CONNECTED, CALL OFFERED (ROUTED), QUEUED (ROUTED), TRUNK SEIZED (NETWORK_REACHED), REJECTED, and ROUTE (REQUEST INSTRUCTION) events when the calling party goes on hold. These flows were uncovered when the STLALTR request was issued after an STLXTND to VDN's which reversed the held and active calls at the vertex station. Symptoms were lost events listed above and associated ERR_UNEXPECTED_CSM_STATE errors logged.

Internal 5985
Description Added more specific error codes when Ethernet link fails to connect.

Added error codes 6428 to 6433 to detect specific TCP/IP connect failures. The new codes look for the conditions EADDRNOTAVAIL, EISCONN, ENETDOWN, ENETUNREACH, and ECONNREFUSED as defined in most TCP/IP communications books. Each new error codes gives tips for how to work around the problem.

Internal T4400
Description Calling party changes after issuing STLRED with AIX Callpath Server.

Fixed internal data structure which was corrupting the calling party info after issuing an STLRED request. Behavior was that the calling party changed to the redirected party. Problem did not occur with the OS/2 or Windows NT servers.

APAR IC20572
Description CONNECT event from station service observe wrongly processed as RECONNECT.

Internal 5997 Fixed processing of CONNECT event generated when a station service observe is initiated against the connected party in an existing two party call where the other party is on hold. This was processed like a RECONNECT and would corrupt the party id field in this call. This is now fixed and the service observer is added to the call via the CONNECT event with the reason STL_ADD_PARTY_RCV_ONLY making a three party call.

APAR IC20177
Description G3 switch never comes up on Windows NT and traps.

The G3 switch will trap at startup if the Callpath Server has not been updated to at least level 1375. Updating the Callpath Server fixes this problem by providing the updated DLL's expected by the G3 code.

Internal 5829
Description Partial support for VDN service observe flows.

This Lucent feature creates a CONNECT event with the connected party as the observer station just prior to the CALL OFFERED event. This corrupted the call state model for this call and subsequent events were discarded. This is now handled by flowing up this unique connect event using a new reason code of STL_ADD_PARTY_RCV_ONLY. Subsequent events are now handling the new call state of TWO_CC_CC and flow as before with the observer party included if applicable. Limitations: If the VDN is a point of entry for external calls and the observer station is monitored, the VDN MUST be monitored also otherwise a call PICKED event will flow. Affected Lucent/CallPath Events: ALERTING, CONNECTED, CALL OFFERED (ROUTED), QUEUED (ROUTED), TRUNK SEIZED (NETWORK_REACHED), REDIRECTED (ROUTED), REJECTED, ROUTE (REQUEST INSTRUCTION). Full support will be offered in a later release.

Internal 5766
Description G3 switch fails to start first time, but does on second try.

Internals 5839,5840 also. Miscellaneous fixes to ensure cleaner shutdown on AIX. Symptom is that the switch dependent code would not start the first time, but the second time it would. The first start actually works, but finds a resource left around from the previous invocation, cleans it up, then immediately exits. Second start always works because it finds no leftover resources.

Internal 5789
Description Only get one FEATURE_INVOKED msg for multi-skilled agent.

Fixed processing of agent LOGIN/LOGOUT events for agents belonging to more than one monitored split. The system has been changed to flow a LOGIN (or LOGOUT) event for each split instead of just the primary skill. The limitation of FEATURE_INVOKED messages for agent state change still applies in that only one will flow per request because this event is generated within the CallPath system and does not come from the G3 switch.

Internal 5759
Description STLINVK requests would be NAK'ed when using WORKGROUP field.

Fixed the utilization of the WORKGROUP field when using STLINVK with ADVANCED_AGENT_ACTIVITY for LOGIN/LOGOUT and AGENT_NOT_READY requests. The reason codes will only be included in the request if the WORKGROUP field is present whereas they were previously always present.

APAR IC19815
Description No response to STLQPS requests for parties of length > 6.

Internal 5728 Fixed STLQPS on parties with length > 6 (external to switch) so it would not leak request tags and always send back a response. Unknown parties are now NAK'ed with INVALID_DIRECTORY_NUMBER and the cause code qualifier is now filled in with the external party number.

Internal 5716
Description Improved error recovery of ISDN link to G3 switch.

Fixed the ISDN link recovery to not shutdown the G3 switch dependent code when a RESTART was sent over the ISDN link or communications timed out due to errors. This change also limits retries to two to three minutes before assuming the cable/connection is broken/unplugged.

Internal 5711
Description REQUEST INSTRUCTION events not handling calls with II digits or flexible billing.

Added support for II digits and Flexible Billing when receiving ROUTE REQUEST's from the switch.

APAR IC19781
Description REQUEST INSTRUCTION events ignored PROGRAM DATA.

Internal 5710 Added support for Program Data when receiving on CALL ORIGINATED events and ROUTE REQUEST's from the switch.

Internal 5680
Description Run out of resources when using three or more CallPath servers to same G3 switch.

Added enhancement to adjust the time when temporary station DOMAIN CONTROLS are released. Set the environment variable MONITORLOOP_TIME=n where 30 <= n <= 600 for the maximum number of seconds the server waits before performing the releases. Usage is helpful for sites where three or more CallPath Servers with local connections to a common G3 switch and the stations are shared among the servers.

Internal 5622
Description Hostname for G3 switch limited to eight characters.

Increased length of hostname for the G3 switch from 8 to 32.

APAR IC19461
Description Get wrong trunk and member data in events.

Internal 5461 Trunk value in both switch trace and application trace always had the same value. Fixed to correctly decode the trunk and member values.

APAR IX70296
Description Restart on error for local G3 connections on AIX does not always work.

Internal 5375 Fixed CallPath Server for AIX with local G3 connection to always recover shared memory, queues, and semaphores after abnormal shutdowns.

APAR IC19191
Description Get unexpected CSM errors using STLRED after a REQUEST INSTRUCTION event.

Internal 5257 Fixed handling of calls generating REQUEST INSTRUCTION events and then redirected using STLRED to other VDN's including the VDN generating the REQUEST INSTRUCTION. These changes also include handling of queuing calls to multiple splits with a queue_to vector steps and then redirecting the call after a wait step.

APAR IC19302
Description Get many CSM errors for predictive dialer applications.

Internal 5215 Fixed many errors for predictive dialing applications. These include correct handling of calls where the external party hangs up before the agent answers and disallowing duplicate MAKECALL requests from the same VDN or hunt group pilot number to the same external party.

APAR IC18891
Description STLRETR after STLXTND to consulting party would fail.

Internal 5214 Fixed negative response to STLRETR when original calling party disconnects in the middle of a consult. This resulted in a lost request tag which could only be recovered by restarting the server.

Internal 5030
Description Extension wrongly removed from call when following cover, no answer path.

Fixed Call Routed event in "cover, no answer" scenario to use the reason STL_RNA_FORWARD_ADD_ROUTE and keep the original alerting station in the call. This allows the original alerting station to correctly connect in the call either manually or via the STLANSR request because the connection id for the extension is no longer deleted by the routed event.

Internal 5029
Description Improved reliability of Ethernet communications to G3 switch.

Added extensive diagnostics into Ethernet communications code for troubleshooting purposes. Added new error codes 6402 to 6426. Set the environment variable ASAI_HBT_TRACE=n where 1 <= n <= 2 and n = 1 is used to trace layer-2 heartbeats and n = 2 for layer-3 heartbeats.

APAR IC18564
Description Trap/core dump when lots of bridged phones ring.

Internal 4956 Fixed trap where more than 16 parties are added to a call using bridged phones. Only occurs when an extension A has the maximum number of bridged appearances (15), and has a coverage path with "cover, no answer" to another extension B that has additional bridged appearances and a call to A does not answer. Changed flow by adding a new Call Routed event with new reason STL_RNA_FORWARD_ADD_ROUTE to the coverage extension B followed by an Alerting event for B and then Alerting events for B's bridged appearances. Only up to 20 parties can be in a call. The error code ERR_TOO_MANY_PARTIES will be logged alerting parties 21 or higher.

APAR IC17991
Description Old calls not cleaned up using CLEAR CALL request.

Internals 4977,5066,5118,5143 Fixed old call cleanup that were stored internally in CSM, but the switch has deleted. Old calls were left around when the external party hung up before a TAKE CONTROL request could be issued to track it or no DISCONNECT/CALL ENDED events were received as a result of a CLEAR CALL request. Problem occurs for outbound predictive makecalls using the options STLSMCP STL_CONNECT_PARTY_TWO_FIRST in conjunction with STLSROP STL_DISCONNECT_REMAINING to clear over dialed calls.

APAR IX70296
Description Core dump generating when AIX Ethernet link goes down.

Internal 5051 Fixed AIX Ethernet TCP/IP cleanup code when SWITCH.EXE is coming down. Resources were not released causing restarts to fail. Csaclean was not finding these resources either and was fixed to find and release them.