|
Problem IC74036
|
Status: Closed |
DB2STOP FORCE MAY HANG IF WRITE SUSPEND IS ISSUED IN PARALLEL |
| product: |
DB2 FOR LUW / DB2FORLUW / 970 - DB2 |
| Problem description: |
A DB2 instance may hang if db2stop force and set write suspend
are issued in parallel.
There exists a timing hole such that a write suspend operation
may successfully suspend a database during a stop force
(deactivate phase). If this happens, the database will enter
write suspend mode, and it will hang.
This may happen in a case where an event monitor attempts to
commit but is unable to because no writing is allowed. The event
monitor then hangs and will refuse to shut down. Here is a stack
of such an example:
<StackTrace>
sqloxconflict_LONG
sqloXLatchNoSpin_track
sqlpxcm1
sqlrrcom_dps
sqlrrcom
sqm_evmon_ttarget::commitInserts
sqm_evmon_ttarget::write_data
sqmEvmonBuffer::write_data
sqmEvmonWriter::flushBuffers
sqmEvmonWriter::cleanupEvmonWriter
sqmEvmonWriter::evmonWriter
sqleIndCoordProcessRequest
sqeAgent::RunEDU
sqlzRunEDU
sqloEDUEntry
</StackTrace> |
| Problem Summary: |
****************************************************************
* USERS AFFECTED: *
* DB2 LUW *
****************************************************************
* PROBLEM DESCRIPTION: *
* The instance is hung when db2stop force is issued. *
* *
* There exists a timing hole such that a write suspend *
* operation may successfully suspend a database during a stop *
* force (deactivate phase). If this happens, the database will *
* enter write suspend mode, and it will hang. *
* *
* This may happen in a case where an event monitor attempts to *
* commit but is unable to because no writing is allowed. The *
* event monitor then hangs and will refuse to shut down. *
* *
* sqloxconflict_LONG *
* sqloXLatchNoSpin_track *
* sqlpxcm1 *
* sqlrrcom_dps *
* sqlrrcom *
* sqm_evmon_ttarget::commitInserts *
* sqm_evmon_ttarget::write_data *
* sqmEvmonBuffer::write_data *
* sqmEvmonWriter::flushBuffers *
* sqmEvmonWriter::cleanupEvmonWriter *
* sqmEvmonWriter::evmonWriter *
* sqleIndCoordProcessRequest *
* sqeAgent::RunEDU *
* sqlzRunEDU *
* sqloEDUEntry *
****************************************************************
* RECOMMENDATION: *
* Upgrade to DB2 v9.7 fp4 or later. *
**************************************************************** |
| Local Fix: |
Do not attempt to stop the instance in parallel with setting
write suspend on a database within the instance. |
| available fix packs: |
DB2 Version 9.7 Fix Pack 4 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 5 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 6 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 7 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 8 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 9 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 9a for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 10 for Linux, UNIX, and Windows
|
| Solution |
db2_v97fp4 will contain the fix for the timing hole. |
| Workaround |
not known / see Local fix |
| Timestamps |
Date - problem reported :
Date - problem closed :
Date - last modified :
| 21.01.2011
08.11.2012
08.11.2012 |
| Problem solved at the following versions (IBM BugInfos) |
9.7.FP4
|
| Problem solved according to the fixlist(s) of the following version(s) |
| 9.7.0.4
|
 |
