Hi, We have a problem with a client where it seems that a VMware backup occurring during a SQLA backup is causing the backup to completely hang with no generated errors in the event log. We have changed the backup time of our own software just now but to get a definitive answer or just to know if anyone else is experiencing any problems would be useful. We had a similar problem with another client where you can see the VMware stats collection service running and coinciding with problems happening with users. One is using SQLA 10.0.1.4310 which is due an upgrade soon. The other is on 16. Is anyone else experiencing problems with backups or general usability on vmware platforms? The server specs are as follows:
|
AFAIK, VMware makes use of the Windows VSS service when doing snapshots of Windows VMs. It seems that those snapshots interfere with SQL Anywhere's own backup. So, when using v16, do you run the SQL Anywhere Volume Shadow Copy Service (VSS) in order to support snapshots of the VM? That tool will issue a CHECKPOINT - and that way prevent (or serialize - I don't know for sure) a parallel backup because that does a CHECKPOINT itself, and checkpoints are serialized. Cf. that other FAQ: How does SQL Anywhere interact with the Windows Volume Shadow Copy Service (VSS)? In case the SQL Anywhere Volume Shadow Copy Service helps to solve your problem with v16, note that it has been introduced with v11, so it is not available with v10. |
We are also facing similar issue. Was the above issue resolved.
Yes, the problem was a conflict with VMware snapshots. Making sure the SQLA VSS service was running and making sure that backup times did not conflict fixed both the stuck backup problem and also the DB freezing when VMware did its snapshot. NB, v11 and above only.