|
Problem IC66274
|
Status: Closed |
INSTANCE MAY TRAP AFTER RESTORE OF EXISTING DATABASE WHEN ROLL FORWARD COMMANDS ARE ISSUED |
| product: |
DB2 FOR LUW / DB2FORLUW / 970 - DB2 |
| Problem description: |
The instance many trap after a restore of an existing database
when a Roll Forward command is issued. The Roll Forward command
may include a check of the status of the Roll Forward as well as
the actual command to start the Roll Forward. The trap is due
to a timing issue during the activation of different processes /
threads.
The the db2loggr process / thread will fail with a signal 11:
2010-01-30-02.12.46.280761+060 E339182911E982 LEVEL: Severe
PID : 17712 TID : 47238155786528 PROC :
db2loggr (SAMPLE) 0
INSTANCE: db2inst1 NODE : 000
FUNCTION: DB2 UDB, oper system services, sqloEDUCodeTrapHandler,
probe:20
DATA #1 : Signal Number Recieved, 4 bytes
11
And will have the following stack:
sqlbInitEDU
sqlpgasn
sqloCreateEDU
sqloRunGDS
sqloInitEDUServices
sqloSystemControllerMain
sqloRunInstance
DB2main
main
The problem is timing dependent, i.e. not always reproducible.
The user can use a non-intrusive db2trc command to change the
timing, thus work around the issue. In the same manner, db2trc
can be used to reproduce the crash. Sample steps to recreate the
crash:
Setup:
======
Create a database. Take an online backup while there is activity
going on the database.
session 1: db2 create db test
session 1: db2 update db cfg for test using logretain yes
session 1: db2 backup db test to /dev/null
session 1: db2 backup db test online
session 1: db2 connect to test
session 1: db2 "create table t1 (c1 char(100))"
session 1: db2 "insert into t1 values ('aaaaaaaaa')"
session 1: while [ 1 ]; do db2 "insert into t1 select * from
t1"; done
session 2: db2 backup db test online
session 1: Ctrl-C
session 1: db2 terminate
Reproduce:
==========
session 1: db2 restore db test
session 1: db2trc on -crash "DB2.SQLP.sqlpgLoggrPostInit.entry"
-suspend
session 1: db2 rollforward db test (this step will hang)
session 2: db2trc off (in order to end the hang)
session 1: (the session will resume, the output will be:)
SQL1224N The database manager is not able to accept new
requests, has
terminated all requests in progress, or has terminated your
particular request
due to a problem with your request. SQLSTATE=55032 |
| Problem Summary: |
****************************************************************
* USERS AFFECTED: *
* Users performing rollforward operation. *
****************************************************************
* PROBLEM DESCRIPTION: *
* The instance many trap after a restore of an *
* existingdatabase when a Roll Forward command is issued. The *
* RollForward command may include a check of the status of *
* theRoll Forward as well as the actual command to start the *
* RollForward. The trap is due to a timing issue during *
* theactivation of different processes /threads.The the *
* db2loggr process / thread will fail with a *
* signal11:2010-01-30-02.12.46.280761+060 E339182911E982 *
* LEVEL:SeverePID : 17712 TID : *
* 47238155786528 PROC :db2loggr (SAMPLE) 0INSTANCE: db2inst1 *
* NODE : 000FUNCTION: DB2 UDB, oper system *
* services,sqloEDUCodeTrapHandler, probe:20DATA #1 : Signal *
* Number Recieved, 4 bytes11And will have the following *
* stack:sqlbInitEDUsqlpgasnsqloCreateEDUsqloRunGDSsqloInitEDUSer
****************************************************************
* RECOMMENDATION: *
* Upgrade to Version 9.7 Fix Pack 2 or later releasecontaining *
* the fix. *
**************************************************************** |
| Local Fix: |
Despite the presence of the "-crash" option of db2trc, these
steps will not cause any crash or other damage to the system.
Instead, a certain code path will merely be put to sleep, and
later resumed.
session 1: db2 restore db test
session 1: db2trc on -crash "DB2.SQLP.sqlpterm.entry" -suspend
session 1: db2 rollforward db test (this step will hang)
session 2: db2trc off (in order to end the hang)
session 1: (the session will resume, the output will be:)
Rollforward Status
Input database alias = test
Number of nodes have returned status = 1
<..etc, the regular success message...> |
| available fix packs: |
DB2 Version 9.7 Fix Pack 2 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 3 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 3a for Linux, UNIX, and Windows
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 9a for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 9 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 10 for Linux, UNIX, and Windows
|
| Solution |
Problem was first fixed in Version 9.7 Fix Pack 2. |
| Workaround |
See Local Fix. |
| Timestamps |
Date - problem reported :
Date - problem closed :
Date - last modified :
| 12.02.2010
25.05.2010
25.05.2010 |
| Problem solved at the following versions (IBM BugInfos) |
9.7.FP2
|
| Problem solved according to the fixlist(s) of the following version(s) |
| 9.7.0.2
|
 |
