SureBackup Errors and Best Practices

[jai64]

I just started using SureBackup. I also just upgrade to 6.5.0.106.

Case# 00152348

(Issue 1)
Since upgrading to version 6.5 my VMs stay open while the complete SureBackup job runs. I did not check: "Keep application group online". Which is weird because that checkbox never worked for me when I checked it in previous versions.

(Issue 2)
I have had a lot of problems with VMs just quitting out with "Results: Cannot detect VM starting because of timeout". Which is weird because I am getting the these errors and the VM backup is fine.

This kills the rest of the SureBackup script from running.

Many things make it quit. No ping, application script error, if there are other devices in the backup that are not in the sure backup like templates. Yet all the VMs are fine and were properly backed up.

1) Question. Support said it is by design. What would be the rational behind this? If I have a backup that does 20 VMS then run SureBackup and one fails, it makes sense to finish the machines to see if there are any more bad devices. Instead, I have to fix one server, then run the SBU again to see if it worked, going through all the servers that worked. I have only been testing for a few days but this is driving me a bit crazy...mostly becuase of the random errors.

2) Question on Best Practice. I have about 75 VMs. I used to have them broken into a bunch of small groups of about 6-10 servers. With the speed increases I swapped it around so I have fewer scripts with more servers in them (25 or 30). Then the SureBackup matches the same machines. Is there a best practice for which is better. Many server in a job or many jobs and few servers.

Thanks Joe.

P.S. How do you get to talk to support these days rather then play email tag with level 1 techs?

[jai64]

After working through this...There are two ways to do a Sure Backup.

1) Via an Application Group.
a) Leaves VMs powered on until Sure Backup job is completed
b) Fails complete Sure Backup job if one VM fails

2) Via the Linked Backup Job.
a) Powers off each VM after it is tested.
b) If a server fails the Sure Backup job it continues testing servers until job is completed.

Note if you have both ways selected in your Sure Backup job it will default to the App Group behavior.

~joe

[Vitaliy S.]

(Issue 2)
I have had a lot of problems with VMs just quitting out with "Results: Cannot detect VM starting because of timeout". Which is weird because I am getting the these errors and the VM backup is fine.

I believe I can explain why you see this message during SureBackup jobs and do not have them while backing up your production VMs. The point is that we boot VMs from the compressed/deduped backup file directly via vPower NFS server. If your target repository is located on either a dedupe device (apart from ExaGrid) or on a low-end hard drive, the process of bringing VMs up (including VMware Tools initialization) might take some time.

1) Question. Support said it is by design. What would be the rational behind this? If I have a backup that does 20 VMS then run SureBackup and one fails, it makes sense to finish the machines to see if there are any more bad devices. Instead, I have to fix one server, then run the SBU again to see if it worked, going through all the servers that worked. I have only been testing for a few days but this is driving me a bit crazy...mostly becuase of the random errors.

Actually, this is true for Application Groups only. If you use linked backup jobs, all your VMs will be processed one by one (the default parameter is to start three VMs at a time). SureBackup job doesn't stop if one of the VMs fails during verification process.

2) Question on Best Practice. I have about 75 VMs. I used to have them broken into a bunch of small groups of about 6-10 servers. With the speed increases I swapped it around so I have fewer scripts with more servers in them (25 or 30). Then the SureBackup matches the same machines. Is there a best practice for which is better. Many server in a job or many jobs and few servers.

I always recommend to group VMs based on their SLA, application or installed OS. Having all VMs in one job might be good for the dedupe, but in this case you will not be able to manage this job in a more granular way.

P.S. How do you get to talk to support these days rather then play email tag with level 1 techs?

You can ask to escalate your support case to a higher tier engineer (if needed) or ask for a call back/WebEx session to verify your configuration/setup.

Hope this helps!

[jai64]

Thanks for the response. I flushed this out with support. I was also able to get a copy of the below doc:

Veeam.v6.x.SureBackup.how-to.and.Troubleshooting.pdf by your tech lead Ben Milligan.

It was a big help and should be added to the user manual.

~joe

[albertwt]

Is there any update or link where can I get the latest PDF ?

[veremin]

Nope, the thread and the documentation mentioned are rather outdated and have not been modified or updated in a while. Do you experience any particular issue with SureBackup configuration or is more like a question of general interest? Thanks!