Very large guest FLR

[albertwt]

Hi All,

Does anyone here ever succeed in restoring single file from very large guest size (eg. > 931 GB) ?

Please share your experience here on how did you set it up and what's the type of the OS.

Thanks.

[TrevorBell]

Hi,

My largest VM is 971 gig. Three weeks ago i had some bad data on one of the drives. to restore the complete backup was reporting 3 days! this was no way near fast enough, so i mounted the image with VB v5 and then copied the files off the mounted VM ( please bear in mind this was v5 beta i was running but it did mount and get back all data ) The production V5 release is alot faster the team have told me.

So my answer is yes i have mounted a large image but i learnt a valuable lesson that day.... its nice to backup and good speed but you need to do period test restores to prove recovery time also.

trev...

[tsightler]

Hi All,
Does anyone here ever succeed in restoring single file from very large guest size (eg. > 931 GB) ?

We have a Windows 2003 VM that is >1.6TB with somewhere upwards of 500,000 files. We end up restoring files on this server several times a month and we've never had any problems just using the normal "Guest files (Windows)" restore method.

The largest Linux VM that we have is 750GB but we've never had a problem with that either using the "Guest files (Other OS)" restore. Takes about 3.5 minutes to bring up the file browser with Veeam 5, but it works, no problem.

Just to test I actually tried using the "Guest files (Other OS)" to restore some files from the 1.6TB Windows host. It took almost 5 minutes for the file browser to come up, but it worked just fine.

So my answer is yes i have mounted a large image but i learnt a valuable lesson that day.... its nice to backup and good speed but you need to do period test restores to prove recovery time also.

While I had learned that lesson many years ago, my discipline had faded a little until our storage array crashed and we lost 4TB+ of VM's and and ~2TB of regular data. Not only did we have to find space for 6TB of data, but then get all of that data restored. We learned a lot about optimizing restore speed that day. It took us ~24 hours to allocate temporary space split between our primary and secondary storage systems and restore all 6TB's of data (not too bad seeing that we're a boot from SAN environment and also lost 6 ESX boot volumes which had to be reinstalled). While it was a true disaster, we learned a lot. We could probably preform the same restore in 6-8 hours now by using techniques we learned from that experience.

Veeam 5 changes this dynamic somewhat, assuming your backup storage together with Veeam Instant Recovery can provide enough IOPS to actually run your systems, but this is a big "if" for servers with even moderate IOP load as the Instant Recovery engine has too much latency for environments that need random IOP performance.

[albertwt]

Thanks guys for the reply, is there any tweak or settings that must be changed in order to perform the FLR from the large VM volume ?
because I always get time out error and the 931 GB disk is not mounted on the FLR appliance.

[tsightler]

I made no changes, it just works. Here are just a few questions I'd try to answer if I was troubleshooting this issue:

How long is it taking before you receive the timeout messages?

Does the FLR VM show the disks attached in vCenter?

Have you watched the FLR VM boot? (I think you answered "yes" to this in another thread)

If you watch the FLR VM boot you should see the Linux log messages scroll by quickly, do you see the disks actually shown in the logs?

Can you ping the FLR VM after it boots?

Does it have an IP on the same subnet as the Veeam server?

[TrevorBell]

Tom.

If you had to restore the complete VM how long would it take you ? Not to pull a few files out of the backup but to actuaully do a full restore from backup of entire VM ? Im just looking at a technology refresh with 10gig iscsi and Dell PS6000XVS just keen to see how long that vm would take you to restore ?

thanks

trev