An overnight job encountered a locking error on updating a record in DB2 10 for z/OS. I've been asked to inspect the DB2 logs to hopefully find more information on what was locking the table. Since I have the timestamp of the failure, how can I look at the DB2 log for a specific time-period?
Thanks for your help!
Dave
If you look in SDSF DA, there should be an entry for *MSTR (e.g., if your DB2 subsystem name is DB2P for your production DB2, then you'd look at DB2PMSTR). If you look at the JESMSGLG of that task, the DB2 logs you are interested will be in there. There should be a DSNT375I (deadlock) or DSNT376I (timeout) entry that will look something like the below (the format is slightly different between the two).
The two error messages will also have a DSNT501I error associated with it that will tell you what the "fight" was over.
12.48.05 STC30099 DSNT376I -DB2P PLAN=DISTSERV WITH 798
798 CORRELATION-ID=program.exe
798 CONNECTION-ID=SERVER
798 LUW-ID=GA5D2A01.CE00.200429174636=79816
798
798 THREAD-INFO=logon_username:server_name:logon_username:program.exe:DYNAMIC:7522055:*:<10
798 .93.42.1.52736.200429174636>
798 IS TIMED OUT. ONE HOLDER OF THE RESOURCE IS PLAN=plan WITH
798 CORRELATION-ID=ENTRabcdefgh
798 CONNECTION-ID=CICSzyxw
798 LUW-ID=USDTCN00.ALDB2P00.D7D81AEAB200=43579
798 THREAD-INFO=CICSzyxw:*:*:ENTRabcd:STATIC:9507687:*:*
798 ON MEMBER DB2P
12.48.05 STC30099 DSNT501I -DB2P DSNILMCL RESOURCE UNAVAILABLE 799
799 CORRELATION-ID=program.exe
799 CONNECTION-ID=SERVER
799 LUW-ID=GA5D2A01.CE00.200429174636=79816
799 REASON 00C9008E
799 TYPE 00000304
799 NAME DBNAME .TBSPACE .X'abcdefg12'.X'11'
In my example, program.exe (logged on as user logon_username from client server_name) was trying to access a page (Identified by X'abcdefg12'.X'11') in database DBNAME, tablespace TBSPACE (the format is determined by the resource type listed in the 501 error code). The page was being held by a CICS task, identified by ENTRabcdefgh in region CICSzyxw.
Hopefully that gets you the information you're looking for.