Exchange server, two jobs, how to truncate logs?

[jpeake]

I have 2 jobs backing up exchange. One that points to onsite storage repo, and the other that points to agent storage in a branch office as target repo.

How should i set log truncation so exchange is fully recoverable from any restore point, from either repo? Or should i not be backing up Exchange in this manner?

[Gostev]

Just choose either job to do this. I would pick the one that backs up locally (as it will run more reliably and finish more consistently).

[zoltank]

I have 2 jobs backing up exchange. One that points to onsite storage repo, and the other that points to agent storage in a branch office as target repo.

How should i set log truncation so exchange is fully recoverable from any restore point, from either repo? Or should i not be backing up Exchange in this manner?

If I'm reading your question correctly, you'll be able to fully recover Exchange from any of the restore points present on either repository.

When the logs are truncated it just means that data is committed to the database. So when the backup runs on the other repository that data is still present, it's just in the database instead of the logs.

[jpeake]

My concern was that if the production DB's get hosed, I need to recover from previous backup and then replay the logs. Exchange keeps the logs even after data is written to the disk database, for this reason. Exchange does not truncate logs after the data is committed to disk. They are truncated only after a successful backup (unless you have circular logging enabled, also know as suicide logging).

I was worried that truncating logs from both backup jobs would create missing logs from the point of view of the other backup. So I think what Gostev said makes sense. Exclude offsite job from truncating, then the log chain is complete from the onsite job's perspective.

Technically I suppose you could truncate from both jobs, but then you might have to look in multiple places to bring back a complete sequential log chain.

[Cokovic]

And i would make sure that the job that is not truncating the logs is running before the one with log truncation on. That way you should have the log files still in both backup locations.

[dellock6]

This is easy to achieve: simply schedule the first one before the second, even if you are not sure when the first ends, one job cannot process a VM while it beeing processed by another job, so the second one will wait in queue for the VM to be released.

Luca.