If you run the following sequence of commands in a logged
database after an L0 archive has been started and before it
finishes and some checkpoints and log switches occur during the
archive and then use this archive to restore the instance
(including the rollforward of the log. logs filled during the
archive), you can see rollforward errors -126 and a chunk being
taken offline during the restore:
alter table <tabname> type (raw)
insert into <tabname>
alter table <tabname> type (standard)
update <tabname>
During the restore, in the rollforward part of logical restore
you will see this type of assertion failure:
14:03:03 Rollforward of log record failed. iserrno = 126
14:03:03 Log Record: log = 15, pos = 0x611c, type =
OLDRSAM:HUPAFT(43), trans = 3619337690157154330
14:03:04 Assert Warning: Chunk 2 is being taken OFFLINE.
14:03:04 IBM Informix Dynamic Server Version 12.10.FC2
14:03:04 Who: Session(14, informix@sles11a, 0, 0x44d2d068)
Thread(43, xchg_1.9, 44cf25b8, 8)
File: rsmirror.c Line: 1926
14:03:04 Results: Dynamic Server will block at next checkpoint
14:03:04 Action: Shutdown (onmode -k) or override (onmode -O)
14:03:04 stack trace for pid 3471 written to
/sqldist/1210fc2/tmp/af.41396f7
14:03:04 See Also: /sqldist/1210fc2/tmp/af.41396f7
14:03:04 Rollforward of log record failed. iserrno = 126
14:03:04 Log Record: log = 15, pos = 0x611c, type =
OLDRSAM:HUPAFT(43), trans = 26
14:03:05 Logical Log 15 Complete, timestamp: 0x2245d4.
14:03:05 Checkpoint blocked by down space, waiting for override
or shutdown
If you have the ONDBSPACEDOWN parameter set to 0 or if you run
'onmode -O' after the instance gets blocked in checkpoint, the
affected chunk will be permanently marked offline. |
