ontape restore and corrupt logical logs

Hi All, I've been called to assist in a restoration of a 7.31.TC7 database. Ontape is used for backups, and logical logs are backed up once a day (!) using "ontape -a", at 5:00am. The system went down at around noon, and would not come up, showing corruption in one of the data chunks and problems with the logical logs. "oncheck -cR" reports lots of header corruptions in the last 2 logical logs. I have a two day old level 0 and a day old level 1 ontape backup, plus backup of logs on disk files (one file per day). I started "ontape -r", which prompted whether I wanted a backup of the still available logical logs, which I wanted. You can select the oldest one, but you cannot select the last you want backed up. What I worry about is that I know that the last two logs are corrupt. There's no way (that I know) to tell ontape to stop at a particular log number, like with onbar. Question is, should I apply this last batch of logs at the end of the restore risking that the corrupt logs will get me back to the start (the full restore takes around 15 hours), or should I be content with restoring only up to 05:00am, and loose half day of data? TIA Gabor

On Thu, 23 Oct 2003 06:46:08 -0400, Gabor Heppes wrote: Contact tech support and have them login and trucate the last two logical logs, then do the restore's logical log backup and restore all of the good logs from files. Art S. Kagel > Hi All, > > I've been called to assist in a restoration of a 7.31.TC7 database. Ontape is > used for backups, and logical logs are backed up once a day (!) using "ontape > -a", at 5:00am. > The system went down at around noon, and would not come up, showing corruption > in one of the data chunks and problems with the logical logs. "oncheck -cR" > reports lots of header corruptions in the last 2 logical logs. I have a two > day old level 0 and a day old level 1 ontape backup, plus backup of logs on > disk files (one file per day). > > I started "ontape -r", which prompted whether I wanted a backup of the still > available logical logs, which I wanted. You can select the oldest one, but you > cannot select the last you want backed up. > > What I worry about is that I know that the last two logs are corrupt. There's > no way (that I know) to tell ontape to stop at a particular log number, like > with onbar. Question is, should I apply this last batch of logs at the end of > the restore risking that the corrupt logs will get me back to the start (the > full restore takes around 15 hours), or should I be content with restoring > only up to 05:00am, and loose half day of data? > > TIA > Gabor

"Gabor Heppes" <gheppes@optushome.com.au> wrote in message news:3f97b174$0$31771$afc38c87@news.optusnet.com.au... > Hi All, > > I've been called to assist in a restoration of a 7.31.TC7 database. Ontape > is used for backups, and logical logs are backed up once a day (!) using > "ontape -a", at 5:00am. > The system went down at around noon, and would not come up, showing > corruption in one of the data chunks and problems with the logical logs. > "oncheck -cR" reports lots of header corruptions in the last 2 logical logs. > I have a two day old level 0 and a day old level 1 ontape backup, plus > backup of logs on disk files (one file per day). > > I started "ontape -r", which prompted whether I wanted a backup of the still > available logical logs, which I wanted. You can select the oldest one, but > you cannot select the last you want backed up. Whilst this is true, you can, by judicious use of dd and block counts, truncate the backed up log archive yourself to contain only those logs known to be good.