1

A little explanation of our setup..

We have an Exchange 2013 in our main site with 1 Database Active on it and then we have a DR Exchange 2013 with the same Database but as passive. Users are currently connecting to both servers since the DR is a connected via a layer 2 bridge and the Main and DR sites are on the same network subnet and also both hosting the CAS and Mailbox roles.

Now the issue is here, when we implemented the DR exchange and created the DAG, we did not setup backup jobs on the DR but left them only on the Main site. We are using Altaro Hyper-V backup. Basically although the VSS writer from Altaro is showing that the backup job is being done on the Main site Exchange, the transaction logs are still accumulating on both servers. Should I setup a backup job on the DR Exchange as well? and if so, should I for example set at 8:00PM a backup job on the Main Exchange and at 1:00am a backup job on the DR Exchange. In this way, would these transaction logs truncate? and is this the best practice please.

StefanGrech
  • 154
  • 2
  • 10
  • Something isn't right here. Altaro may be able to backup from either the active or the passive database copies (I don't know which it does, you'll need to read the documentation) but there is only one set of transaction logs per mailbox database, regardless of how many copies of that database there are. The logs are copied from the active copy to the passive copy and then replayed into the passive copy. If the logs aren't being flushed after the backup then that's why they're accumulating. You need to find out why the backup isn't flushing the transaction logs. – joeqwerty Dec 29 '15 at 23:30

2 Answers2

1

Basically the issue was found to be because there was some minutes difference in the local server time on both servers.. After the time was synced with the local AD, the issue seems to be resolved and also the logs have been truncated successfully.

Thanks for your Help ;)

StefanGrech
  • 154
  • 2
  • 10
0

Logs continue to build for one of two reasons in my experience.

  1. The DAG isn't behaving itself, so they are out of sync or other problem.
  2. The Backup isn't actually working correctly.

The first thing to do is check the properties of the database. See whether the backup flag has actually been set with the correct date and time. If not, then that is the problem.

If the backup flag is set, then look at your DAG to see that it is in sync and both copies are healthy.

Another trick I have used in the past is to do a backup with another product. You can download the trial of Backup Assist for free and run a backup using that. That would give you a valid backup.

Have you checked that the version of the backup software you are using supports the version of Exchange and Windows? In the configuration there could be an option to do a full or copy backup - full is what will clear the logs.

Simon.

Sembee
  • 2,884
  • 1
  • 8
  • 11
  • Basically as I said, we are taking backups only on the Main site Exchange where the database is active.. on the other server we've set the database to not go active automatically (blocked).. Last night we took a backup successfully but when I checked the last backup time on the exchange ECP , it is still showing as a much earlier date.. Also I noticed that the Exchange VM rebooted automatically as the backup finished.. Could it be that it was going to purge the logs btut for some reason failed to do so on the DR exchange and caused this restart ? – StefanGrech Dec 31 '15 at 08:04
  • The lack of change on the backup value on the mailbox is why the logs aren't flushing - Exchange doesn't know that the backup has completed. If the Exchange server is rebooting, then something is obviously going wrong. As the backup tool you are using has its own VSS writer, rather than using the native one, the first thing I would do is remove the backup tool, reboot the server to flush it out completely, then reinstall it - checking you have the latest build from the backup vendor. I would also check whether anything is logged for the reboot reason. – Sembee Dec 31 '15 at 09:24
  • Exchange does have some measures within that will trigger a reboot as part of its health monitoring. – Sembee Dec 31 '15 at 09:24