![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
Hi there, New to DB/2, but have been DBA'ing on Ingres for some time. We're running DB/2 9.7 Fix Pack 1. We have a Cognos BI database which began refusing connections with the 58030 error. It turns out our log disk ran out of space, and the logs were moved to another location. We then moved the logs back to the correct location and tried reconnecting. Unfortunately we now are getting the SQL1042C/58004 error, and it looks like the first of the logs has a problem. We seem to have hit a perfect storm with this. The log (3538) is corrupt, and dates back to 19/3/2012. We have an online database backup from yesterday which was run excluding logs (but db2ckbkp shows that the first referenced log file is the corrupted 3538 from two days earlier). The "delete old backups" script has run overnight and removed any old backups, and from what I can gather no successful backup to tape has happened in the relevant window. Are we now basically toast, and are looking to try recreate this database ? If it were Ingres, I'd be able to do a rollforward and exclude the journal files to restore to the point in time of the backup without any further processing - is this possible in DB/2 with an online backup ? From my frantic reading this morning it looks like it would be possible with an offline backup, but restoring an online backup without further processing basically dumps me back to where I currently am ? Thankfully it's a Cognos BI database so all the source data is available, but it's 69Gb and I'd rather avoid recreating it from scratch if possible... Thanks, - Tony |
#3
| |||
| |||
|
|
Hi there, New to DB/2, but have been DBA'ing on Ingres for some time. We're running DB/2 9.7 Fix Pack 1. We have a Cognos BI database which began refusing connections with the 58030 error. It turns out our log disk ran out of space, and the logs were moved to another location. We then moved the logs back to the correct location and tried reconnecting. Unfortunately we now are getting the SQL1042C/58004 error, and it looks like the first of the logs has a problem. We seem to have hit a perfect storm with this. The log (3538) is corrupt, and dates back to 19/3/2012. We have an online database backup from yesterday which was run excluding logs (but db2ckbkp shows that the first referenced log file is the corrupted 3538 from two days earlier). The "delete old backups" script has run overnight and removed any old backups, and from what I can gather no successful backup to tape has happened in the relevant window. Are we now basically toast, and are looking to try recreate this database ? If it were Ingres, I'd be able to do a rollforward and exclude the journal files to restore to the point in time of the backup without any further processing - is this possible in DB/2 with an online backup ? From my frantic reading this morning it looks like it would be possible with an offline backup, but restoring an online backup without further processing basically dumps me back to where I currently am ? Thankfully it's a Cognos BI database so all the source data is available, but it's 69Gb and I'd rather avoid recreating it from scratch if possible... Thanks, - Tony I would recommend that you open up a PMR with IBM support. They are in |
![]() |
| Thread Tools | |
| Display Modes | |
| |