Kbase P138816: Replication Agent crashes when rolling forward busy AI extent to target database
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
17/03/2009 |
|
Status: Unverified
SYMPTOM(s):
Replication Agent crashes when rolling forward busy AI extent to target database
Last (busy) AI extent is called (e.g.): sports2000.a7Last archived AI file is called (e.g.): sports2000.20081217.035906.00000006.sports2000.a6
DSRUTIL target -C applyextent tries to apply an extent named as follows:sports2000.20081217.035906.00000006.sports2000.a6
DSRUTIL seems to hang and needs to be terminated with CONTROL+C
Stack trace from rpagent.exe reads:
//=====================================================PROGRESS stack trace as of Wed Dec 17 03:33:31 2008//=====================================================Exception code: C0000005 ACCESS_VIOLATIONFault address: 0040123B 01:0000023B C:\PROGRESS\OPENEDGE\10.1C\BIN\RPAGENT.EXE
Registers:EAX:00000000EBX:00DEC727ECX:776D6D2EEDX:00000001ESI:003CB110EDI:003C5290CS:EIP:001B:0040123BSS:ESP:0023:0012FD00 EBP:003C5290DS:0023 ES:0023 FS:0038 GS:0000Flags:00010246
Call Stack:Address Frame0040123B 003C5290 0001:0000023B C:\PROGRESS\OPENEDGE\10.1C\BIN\RPAGENT.EXE003CB110 00DEBCAF 0000:00000000
FACT(s) (Environment):
Target DB is down, activity on source causes further AI switches
AI Management utility / aiarchiver is being used
AI files that had already been backed up and re-named by the archiver can be rolled forward
Windows
OpenEdge 10.1C02 Service Pack
OpenEdge Replication 10.1C02 32-bit Service Pack
CAUSE:
Bug# OE00179020
This
crash seems to occur due to the fact that the name of the last (busy)
AI extent is shorter than the name of the archived AI extents.
FIX:
There is no fix at this moment in time.
A workaround is to copy and rename the last database AI extent (in the example above that would be sports2000.a7) and rename it so its name contains as many characters as the archived AI extents contain.