Thread 1 cannot allocate new log

Listers,

What's the best approach to deal with the archiver not keeping up with the redo log switches? Currently, redo logs are 20 MB, originally had four groups. We have one process that runs twice a day and results in one or two 'cannot allocate new log' errors. The switch occurs seconds later without causing the update to fail, and the process completes in less than 10 minutes. Tried increasing the size of the logs, no improvement. Added two more log groups, errors stopped for a time, but then reoccurred. This process populates a denormalized table for queries, and the problem began after additional data was added to the base tables. Typically, I see only 4 to 12 switches within a 24 hour period, so I'm not sure if I should continue to added more groups or increase the size. If I increase the size, should I worry about checkpoints occurring too infrequently? (Only occur at log switches now). Both copies of the redo logs and the archived redo logs are all on separate disks.

Robyn

Lend Lease REI Received on Mon Jun 12 2000 - 09:15:32 CDT