PQ61089: WEBSPHERE SERVER REGIONS REMAIN WHEN SERVER IS SHUTDOWN OR RECYCLED

 A fix may be available

Obtain the fix for this APAR



APAR status
Closed as program error.

Error description
When the stop command is issued to normally shut down an
application server, either the control region comes down
normally but its corresponding server region(s) remain up
or neither the control nor server region(s) end.  Another
scenario is where the Systems Management Server internally
issues a stop to an application server when that server
was running and a conversation containing a change to that
application server was activated with the Administration
Application (SM EUI), where the activation never completes
because the application control and/or server region(s)
would not complete normal termination.  Another scenario
where this problem can be seen is when a server region is
recycled, such as for garbage collection, where a new
server region is started and will accept work but the
recycled server hangs in normal termination.
Local fix
Manually cancel the application server's control and/or
server region(s).
Problem summary
****************************************************************
* USERS AFFECTED: All users of WebSphere Application Server    *
*                 V4.0.1 for z/OS and OS/390                   *
****************************************************************
* PROBLEM DESCRIPTION: When a stop command is issued to        *
*                      normally shut down an application       *
*                      server, either the control region comes *
*                      down normally but its corresponding     *
*                      server region(s) remain up or neither   *
*                      the control nor server region(s) end.   *
****************************************************************
* RECOMMENDATION:                                              *
****************************************************************
Stop server brought down the control region but not the
server region. For this case a dump of the server region shows
the server region main thread called destroyJavaEnv which called
jni_DestroyJavaVM who does not return because it is waiting,
appearing to hang.

Stop server did not bring down the control region or server
region(s). For this case, a dump of the control and server
region shows a control region thread in deactivate_impl which
called bboosstp which is waiting on the boam_stop_srr_ecb. The
server region threads are in their normal state of waiting
for work from WLM. The post was skipped because the queue
did not look empty to the server region. The elements
were marked deleted but not removed from the queue
when the server region released the shared lock because
the control region still held the shared lock.
Problem conclusion
For the case where the server region call to destroyJavaEnv
is not completing, provide an environment variable
BBOO_SKIP_DESTROYJAVAENV. When BBOO_SKIP_DESTROYJAVAENV=YES
is specified, bbooboai.cpp will be updated to not call
destroyJavaEnv.

If your installation experiences server regions not ending once
a Stop command is issued for the server, the following steps
should be followed to relieve this problem:

1) Add an environment variable using the Administration
   Application, SM EUI. This should be done by adding a
   conversation to the SM EUI.  Either in the list of
   environment variables for the Sysplex properties or
   the list of environment variables in the properties for each
   individual server, J2EE servers as well as System Management,
   Naming, and IR, add an environment variable named
   BBOO_SKIP_DESTROYJAVAENV and give it a value of YES.

   Note: Setting the environment variable
         BBOO_SKIP_DESTROYJAVAENV to a value of YES on the
         Sysplex object in the Administration Application
         (SM EUI) will have it be in affect for all servers.
         It will also be in affect for any new servers defined
         in the future when set at the Sysplex object level.

2) Commit and activate the conversation.  Recycle the WebSphere
   runtime so that the new setting is picked up for System
   Management, Naming, and IR servers.  Also recycle any
   application servers to force the value to be picked up. The
   new setting will also be picked up when any new server
   regions start up after the activate of the conversation that
   added the environment variable.

For the case where the control region and server regions did not
come down, bboosstp.plx has been updated. After freeing the
shared lock, bboosstp.plx has been updated to not wait if the
server region registry is empty.

APAR PQ61089 is associated with SERVICE LEVEL W401070 of
WebSphere Application Server V4.0.1 for z/OS and OS/390.
Temporary fix Comments
APAR information
APAR number PQ61089
Reported component name WASKBASE
Reported component ID 5655A9801
Reported release 401
Status CLOSED PER
PE NoPE
HIPER NoHIPER
Submitted date 2002-05-13
Closed date 2002-05-31
Last modified date 2002-07-03

APAR is sysrouted FROM one or more of the following:

APAR is sysrouted TO one or more of the following:
UQ66834

Modules/Macros
BBOOBOAI BBOOSSTP BBOUBINF      

Fix information
Fixed component name WASKBASE
Fixed component ID 5655A9801

Applicable component levels
R401 PSY UQ66834    UP02/06/06 P F206

  Fix is available
Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.


Document Information


Product categories: Software > Application Servers > Distributed Application & Web Servers > WebSphere Application Server for z/OS
Operating system(s):
Software version: 401
Software edition:
Reference #: PQ61089
IBM Group: Software Group
Modified date: Jul 3, 2002