Help Instant Restore running need to get exchange back

[ober72]

Hi Folks,

I am dealing with support but also in need of advice. Case number 01062274

We had to do an instant restore of a large Exchange server when we lost a drive a couple of days ago. We tried a migrate to production but there was a mistake with space allocation and the job failed. Then Veeam support told us instead to do a replication job off of the instant restore. 3 disks were completed but the fourth was taking foreever and we found it was using a proxy in a different location. So we stopped the job disabled the proxy. The job started much faster but started to affect production and we were forced to stop it. Now we are thinking about simply getting the logs and edb off of the running Instant restore and then placing them in the original lost disk.

Can someone advise what they think would be the best method to proceed as speedily as possible The exchange server is functioning in the Instant restore but it has now been a number of days.

thanks

[foggy]

How about svMotion? What kind of space allocation error did you have when attempted to migrate the VM? Any chance to perform replication using the right proxy during after-hours, so that not to impact production? What kind of impact it was, btw?

[ober72]

we could try that, but I think it’s because the veeam server is running out of resources. What happens when we run the replication, is the Exchange server becomes unresponsive. What if we shut the VM Down can we then do a direct recovery on the files from the instant VM?

Support told us to do a replication job and not migrate as well.

[ober72]

I mean shutting down the exchange services and not the vm in the instant restore of course as then I assume everything would vanish, i.e. all the changes from the last 3 days

[foggy]

we could try that, but I think it’s because the veeam server is running out of resources.

What specifically is because of that? Could you describe your setup in a bit more detail?

What happens when we run the replication, is the Exchange server becomes unresponsive.

Short stun might be expected on snapshot commit phase.

[ober72]

Not sure, but everything was fine but just very slow for one disk when we were using the other proxy by accident in a remote location. We now have another powerful server on the same switch next to the veeam server, if we use that as the only proxy then would that then unload work from the Veeam server.

[ober72]

What if we created a special vm and then used virtual appliance mode?

[foggy]

If some other proxy server is used to perform data processing (not the default one, installed as a part of Veeam B&R installation), then backup server is not loaded at all. In case the link is slow, make sure you have proxies on both sides. Yes, you can try hotadd proxies as well.

[ober72]

Hi Foggy,

We have it running now, one question can we change the backup proxy setting on the fly i.e. to turn on throttling and turn it back off without affecting the replication job?

Also If we shutdown the windows exchange server inside of the instant restore (creating some downtime for us) will this speed up the replication job and will the instant restore job still be running without issues, i.e. could we then boot up the server from within the instant restore and be where we were when we just it down?

Thanks

[foggy]

We have it running now, one question can we change the backup proxy setting on the fly i.e. to turn on throttling and turn it back off without affecting the replication job?

Talking specifically about throttling, these settings can be changed on the fly and are applied immediately.

Also If we shutdown the windows exchange server inside of the instant restore (creating some downtime for us) will this speed up the replication job and will the instant restore job still be running without issues, i.e. could we then boot up the server from within the instant restore and be where we were when we just it down?

Yes, you can shutdown the instantly recovered VM and boot it up later, changes will be preserved. This will result in less changes performed inside guest OS.