Oninit cdi / Archive
Talk to an expert →

how to restore a dbspace that is NOT down...

Posted in 2012

Topics: Backup & Restore, Storage & Space Management, Error Codes & Troubleshooting, Logging & Checkpoints, Versions, Editions & End-of-Life

Kern Doe — — source: IIUG Forums & Mailing Lists

Chunks that associate with the dbspace are NOT down but one of them is in inconsistent state, I tried to restore just the dbspace by: onbar -r dbsname1 but it says 2012-11-09 15:27:45 20223 20215 (-43394) Storage space dbsname1 is not down so it will not be restored. 2012-11-09 15:27:45 20223 20215 There are no storage spaces/logical logs to backup/restore. So how can we still restore...? 11/09/12 14:50:47 Assert Failed: Page Check Error in chk_ext_cf:bad chunk free list page 11/09/12 14:50:47 IBM Informix Dynamic Server Version 11.50.FC9W1 11/09/12 14:50:47 Who: Session(202601, informix@p1, 13646, 2876575c8) Thread(2382839, oncheck, 287610708, 1) File: rsdebug.c Line: 1121 11/09/12 14:50:47 Results: Possible inconsistencies in a Chunk Freelist page 11/09/12 14:50:47 Action: Run 'oncheck -ce' and/or restore affected DBSpace 11/09/12 14:50:47 stack trace for pid 15317 written to /f1/dumpdir/af.5fdf6ca6 11/09/12 14:50:47 See Also: /f1/dumpdir/af.5fdf6ca6, shmem.5fdf6ca6.0 11/09/12 14:50:58 Error writing '/f1/dumpdir/shmem.5fdf6ca6.0' errno = 28 11/09/12 14:50:58 Page Check Error in chk_ext_cf:bad chunk free list page

↪ replying to Kern Doe
George_Palmer@aotx.uscourts.gov — — source: IIUG Forums & Mailing Lists

Kern, 1) Change you DUMPDIR to a volume or directory that will hold a file the size of the shared memory. so you can get the dump ( or change DUMPSHMEM to 0 ) to avoid that extra issue in the issue. 2) Are you using raw disk or cooked file ? I have had issue where a chunk added to a dbspace that is actually already part of a different instance will produce an error similar to this. When a chunk is added and is raw space, only a single page is updated, The chunk free list ( provided this is not the 1st chunk in a dbspace. 3) Did you attempt the recommended oncheck -ce ? ( from the error message ) this can fix some issues. 4) If you are really attempting a restore I believe the onspaces -s might be able to be used to set the chunk down ( are you sure you would want to do this ) then the restore could be attempted. Also it might be interesting to run an oncheck -pe ( which may show where the free page list rows are in error ( if it still works ) George. From: "Kern Doe" <kern_doe@yahoo.com> To: ids@iiug.org Date: 11/09/2012 03:40 PM Subject: how to restore a dbspace that is NOT down... [28781] Sent by: ids-bounces@iiug.org Chunks that associate with the dbspace are NOT down but one of them is in inconsistent state, I tried to restore just the dbspace by: onbar -r dbsname1 but it says 2012-11-09 15:27:45 20223 20215 (-43394) Storage space dbsname1 is not down so it will not be restored. 2012-11-09 15:27:45 20223 20215 There are no storage spaces/logical logs to backup/restore. So how can we still restore...? 11/09/12 14:50:47 Assert Failed: Page Check Error in chk_ext_cf:bad chunk free list page 11/09/12 14:50:47 IBM Informix Dynamic Server Version 11.50.FC9W1 11/09/12 14:50:47 Who: Session(202601, informix@p1, 13646, 2876575c8) Thread(2382839, oncheck, 287610708, 1) File: rsdebug.c Line: 1121 11/09/12 14:50:47 Results: Possible inconsistencies in a Chunk Freelist page 11/09/12 14:50:47 Action: Run 'oncheck -ce' and/or restore affected DBSpace 11/09/12 14:50:47 stack trace for pid 15317 written to /f1/dumpdir/af.5fdf6ca6 11/09/12 14:50:47 See Also: /f1/dumpdir/af.5fdf6ca6, shmem.5fdf6ca6.0 11/09/12 14:50:58 Error writing '/f1/dumpdir/shmem.5fdf6ca6.0' errno = 28 11/09/12 14:50:58 Page Check Error in chk_ext_cf:bad chunk free list page ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.

↪ replying to George_Palmer@aotx.uscourts.gov
Kern Doe — — source: IIUG Forums & Mailing Lists

George, Thank you but I read your email a little too late. I wish I could try all these valuable steps from yours. Anyway, we called IBM Informix, they said they may have to dial in, too much of an issue, and we don't know how that worked here having someone to dial in. So we decided with IBM support to restore the whole system. We shutdown the instance and issued "point-in-time" restore. It's been sitting for a while now, don't know what's happening... ... 2012-11-09 16:25:21 12317 12315 Found a L1 backup object without an earlier L0. Removing it from the timeline. mnw_ctn_dbs381 ND 2012-01-27 04:27:04 2012-11-09 16:25:43 12317 12315 Successfully connected to Storage Manager. 2012-11-09 16:26:06 12317 12315 Begin reserved pages restore (level 0 of rootdbs, Storage Manager copy ID: 1 1351913638). 2012-11-09 16:26:16 12317 12315 Completed reserved pages restore (level 0 of rootdbs). 2012-11-09 16:55:19 12317 12315 Successfully connected to Storage Manager. 2012-11-09 17:07:19 12317 12315 (-43296) WARNING: BAR_TIMEOUT Storage Manager Progress may be stalled. 2012-11-09 17:47:21 12317 12315 (-43296) WARNING: BAR_TIMEOUT Storage Manager Progress may be stalled.. ________________________________ From: "George_Palmer@aotx.uscourts.gov" <George_Palmer@aotx.uscourts.gov> To: ids@iiug.org Sent: Friday, November 9, 2012 4:13 PM Subject: Re: how to restore a dbspace that is NOT down... [28782] Kern, 1) Change you DUMPDIR to a volume or directory that will hold a file the size of the shared memory. so you can get the dump ( or change DUMPSHMEM to 0 ) to avoid that extra issue in the issue. 2) Are you using raw disk or cooked file ? I have had issue where a chunk added to a dbspace that is actually already part of a different instance will produce an error similar to this. When a chunk is added and is raw space, only a single page is updated, The chunk free list ( provided this is not the 1st chunk in a dbspace. 3) Did you attempt the recommended oncheck -ce ? ( from the error message ) this can fix some issues. 4) If you are really attempting a restore I believe the onspaces -s might be able to be used to set the chunk down ( are you sure you would want to do this ) then the restore could be attempted. Also it might be interesting to run an oncheck -pe ( which may show where the free page list rows are in error ( if it still works ) George. From: "Kern Doe" <kern_doe@yahoo.com> To: ids@iiug.org Date: 11/09/2012 03:40 PM Subject: how to restore a dbspace that is NOT down... [28781] Sent by: ids-bounces@iiug.org Chunks that associate with the dbspace are NOT down but one of them is in inconsistent state, I tried to restore just the dbspace by: onbar -r dbsname1 but it says 2012-11-09 15:27:45 20223 20215 (-43394) Storage space dbsname1 is not down so it will not be restored. 2012-11-09 15:27:45 20223 20215 There are no storage spaces/logical logs to backup/restore. So how can we still restore...? 11/09/12 14:50:47 Assert Failed: Page Check Error in chk_ext_cf:bad chunk free list page 11/09/12 14:50:47 IBM Informix Dynamic Server Version 11.50.FC9W1 11/09/12 14:50:47 Who: Session(202601, informix@p1, 13646, 2876575c8) Thread(2382839, oncheck, 287610708, 1) File: rsdebug.c Line: 1121 11/09/12 14:50:47 Results: Possible inconsistencies in a Chunk Freelist page 11/09/12 14:50:47 Action: Run 'oncheck -ce' and/or restore affected DBSpace 11/09/12 14:50:47 stack trace for pid 15317 written to /f1/dumpdir/af.5fdf6ca6 11/09/12 14:50:47 See Also: /f1/dumpdir/af.5fdf6ca6, shmem.5fdf6ca6.0 11/09/12 14:50:58 Error writing '/f1/dumpdir/shmem.5fdf6ca6.0' errno = 28 11/09/12 14:50:58 Page Check Error in chk_ext_cf:bad chunk free list page ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum. ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.

Glossary