Managing large log tapes

Client has 80GB tape drives and wants to use continuous log backup, with ontape for the time being. Although he will also be doing daily level 0 archives, so he shouldn't have massive rollforwards, he is concerned that if he has to roll forward while the log tape had got near to its end, it could take IDS a long time to locate which logs it needs to roll forward. Does anyone have any sagely advice they could add to this situation? -- -Andy Kent- Bristol, England Remove '.DontSpamMe' from return address to respond.

The purpose of the logical log backup is to restore changes since the last level-backup. Therefore, once to did a backup (example level-0), and it's successful, you no longer need your logical log backup. Follow the steps before doing a backup (assuming continuous logical backup is running): 1. run onmode -l, then onmode -c, to force to the next log file, hence, it will write the last log to the continuous log backup. 2. stop the continuous log backup (press control-c or kill the process) 3. remove the continuous log backup tape (this is part of the backup set). Replace it with a new tape for the new continuous log backup 4. run the new continuous log backup (ontape -c) 5. do a level backup (level 0 , 1, 2). After this, the previous logical log backup is no longer needed. Also he doesn't have to do a level 0 daily. Tell him that he can also use level-1, then level-2, so that the backup time is not going to be the same as level-0. In article <oJOX5.700$lw2.11861@news6-win.server.ntlworld.com>, "Andy Kent" <andykent.bristol.DontSpamMe@virgin.net> wrote: > Client has 80GB tape drives and wants to use continuous log backup, with > ontape for the time being. Although he will also be doing daily level 0 > archives, so he shouldn't have massive rollforwards, he is concerned that if > he has to roll forward while the log tape had got near to its end, it could > take IDS a long time to locate which logs it needs to roll forward. > > Does anyone have any sagely advice they could add to this situation? > > -- > -Andy Kent- > Bristol, England > > Remove '.DontSpamMe' from return address to respond. > > Sent via Deja.com http://www.deja.com/ Before you buy.

Andy Kent wrote: > > Client has 80GB tape drives and wants to use continuous log backup, with > ontape for the time being. Although he will also be doing daily level 0 > archives, so he shouldn't have massive rollforwards, he is concerned that if > he has to roll forward while the log tape had got near to its end, it could > take IDS a long time to locate which logs it needs to roll forward. > > Does anyone have any sagely advice they could add to this situation? > Change the logical log tape every time a level-0 archive is performed. Brett Randall

Andy Kent wrote in message ... >Client has 80GB tape drives and wants to use continuous log backup, with >ontape for the time being. Although he will also be doing daily level 0 >archives, so he shouldn't have massive rollforwards, he is concerned that if >he has to roll forward while the log tape had got near to its end, it could >take IDS a long time to locate which logs it needs to roll forward. > >Does anyone have any sagely advice they could add to this situation? You don't need to swap log tapes at the same time as any archive, unless of course you are sharing the same tape drive. Log tapes only need removing and carefully storing away when: 1) You stop the continuous logs 2) That can include bringing down the engine - of course the continuous logs stop then! 3) The engine craps out - ditto reason. 4) The size of the contents becomes too large for rollbacks in the event of a needed recovery. IF, during a restore, it prompts for a particular log record, say 2100, and you have a tape that starts with log 2000, then it will skip forward to the position of log 2100. The only issue to consider is the time it takes to run thru the logs (including the ones it skips over) vs. the convenience or otherwise of changing them at a suitable schedule. The time it takes to skip could be easily determined by the size of the logs written, and how many get written per day/week/month. Most important of all: write down the numbers of the logs (also the log numbers mentioned by the archives) in a nice neat book with columns, so that you have excellent chances of successful recovery. Having "sufficient" physical and logical log spaces is also a good idea... A prerequisite for any archiving activities, especially on a 24x7 engine. I wouldn't be bothering to force checkpoints, force changes to next logical log, or force the termination of continuous logging prior to commencement of archiving. It only makes the ritual more complicated for the clerically-minded individual you assign to the task, and will not improve your convenience or chances of successful recovery in the event of a restore.