|
Problem IC76273
|
Status: Closed |
JAVA UDF RUNNING IN FENCED MODE ON WINDOWS CAN ERROR OUT FOLLOWED BY MESSAGE "MARKING FMP AS UNSTABLE" IN DB2DIAG.LOG. |
| product: |
DB2 FOR LUW / DB2FORLUW / 970 - DB2 |
| Problem description: |
Java user defined function (UDF) running in fenced mode on
windows can in some cases error out because of a timing hole
issue caused by synchronization problem between the db2fmp
process and db2agent process. This problem only occurs on
Windows.
In the db2diag.log you will have errors looking like this
2010-09-10-12.03.31.056000+060 I41631739H492 LEVEL: Severe
PID : 3616 TID : 2540 PROC :
db2syscs.exe
INSTANCE: DB2 NODE : 000 DB : SCORED
APPHDL : 0-29 APPID:
10.44.66.72.8967.100910105558
AUTHID : ANADMIN
FUNCTION: DB2 UDB, routine_infrastructure, sqlerReturnFmpToPool,
probe:900
MESSAGE : marking fmp as unstable:
DATA #1 : Hexdump, 4 bytes
0x6061C270 : 9C0A 0000 ....
and it's followed by other errors that imply that the fenced
mode processing environment of DB2 have stopped to work.
Take note! that the symptoms above can also point to that your
java udf contain a coding issue or that your system is running
out of some resource or there been some system error. Only in
rare cases you will hit this particular problem. You're more
likely to hit this problem on a system that is heavily loaded
and run slow.
Other possible reasons for the error above are:
Running a java stored procedure that is running out of java
heap.
The db2fmp process is running Java routines, and the java heap
is full in the existing db2fmp process.
The db2fmp process is being forced by a user that forced the
application.
When the stored procedure itself has a trap within it's own
code, not necessarily within DB2 code.
The db2fmp process is having problem starting threads or
creating semaphores because of system resource allocation /
consumption issue, what may cause SQL1042 or SQL1131 being
returned.
The db2fmp process encounters communication errors from
unexpected system error. |
| Problem Summary: |
****************************************************************
* USERS AFFECTED: *
* Users that use java udfs in fenced mode. *
****************************************************************
* PROBLEM DESCRIPTION: *
* Java user defined function (UDF) running in fenced mode on *
* windows can in some cases error out because of a timing hole *
* issue caused by synchronization problem between the db2fmp *
* process and db2agent process. This problem only occurs on *
* Windows. *
* *
* In the db2diag.log you will have errors looking like this *
* *
* 2010-09-10-12.03.31.056000+060 I41631739H492 LEVEL: *
* Severe *
* PID : 3616 TID : 2540 PROC : *
* db2syscs.exe *
* INSTANCE: DB2 NODE : 000 DB : *
* SCORED *
* APPHDL : 0-29 APPID: *
* 10.44.66.72.8967.100910105558 *
* AUTHID : ANADMIN *
* FUNCTION: DB2 UDB, routine_infrastructure, *
* sqlerReturnFmpToPool, *
* probe:900 *
* MESSAGE : marking fmp as unstable: *
* DATA #1 : Hexdump, 4 bytes *
* 0x6061C270 : 9C0A 0000 .... *
* *
* and it's followed by other errors that imply that the fenced *
* mode processing environment of DB2 have stopped to work. *
* *
* Take note! that the symptoms above can also point to that *
* your *
* java udf contain a coding issue or that your system is *
* running *
* out of some resource or there been some system error. Only *
* in *
* rare cases you will hit this particular problem. You're more *
* likely to hit this problem on a system that is heavily *
* loaded *
* and run slow. *
* *
* Other possible reasons for the error above are: *
* *
* Running a java stored procedure that is running out of java *
* heap. *
* *
* The db2fmp process is running Java routines, and the java *
* heap *
* is full in the existing db2fmp process. *
* *
* The db2fmp process is being forced by a user that forced the *
* application. *
* *
* When the stored procedure itself has a trap within it's own *
* code, not necessarily within DB2 code. *
* *
* The db2fmp process is having problem starting threads or *
* creating semaphores because of system resource allocation / *
* consumption issue, what may cause SQL1042 or SQL1131 being *
* returned. *
* *
* The db2fmp process encounters communication errors from *
* unexpected system error. *
****************************************************************
* RECOMMENDATION: *
* Upgrade to fp5 or later. *
**************************************************************** |
| Local Fix: |
Performance tuning of DB2 instance might help. |
| available fix packs: |
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 |
Fixed in DB2 v9.7 fp5. |
| Workaround |
not known / see Local fix |
| Timestamps |
Date - problem reported :
Date - problem closed :
Date - last modified :
| 09.05.2011
20.12.2011
20.12.2011 |
| Problem solved at the following versions (IBM BugInfos) |
9.7.FP5
|
| Problem solved according to the fixlist(s) of the following version(s) |
| 9.7.0.5
|
 |
