Health Check on Remote Backup Copy Job

[redhorse]

Hello,

we have the following configuration:

Source job:

• Incremental with weekly synthetic full backups

• Keep 4 weekly

• daily surebackup

• weekly health check

Backup copy job:

• Immediate copy

• 14 days retention and keep 4 weekly and 12 monthly

• Hardened immuatable repository at remote side

• disabled health check

We had initially enabled health check and it took more than one day, there is a direct connection between both sides with 400 MBit/s and it's about 12 TB data. We'd like to enable health check for the backup copy repository but I am still not sure how to achieve this without having it running for days. Actually there is no chance to install a server at the remote side. Maybe there is another way to do this? In the GUI it's not possible to split the health check in groups, or did I miss something?

Just for my clarification, the health check checks the latest restore point but with the whole chain (VBK+VIBs) to ensure that this restore point is valid for restore. Am I right? The documentation is somehow not clear for me as it sounds like that only the latest restore point (just the VIB) is checked.

Thanks!

[david.domask]

Hi redhorse,

Just a quick question, is the Mount Server located on the source site? I'm guessing that's why it's taking a bit for the processing as the mount is happening cross sites. You mentioned there's no chance to put a server on the remote site, but is there a server that could act as mount server?

And yes, by default SureBackup checks the most recent point in the full verification mode, but you can tell Surebackup to check specific restore points if desired. So the most recent point is verified in the full verification mode to ensure it can boot and pass the SureBackup boot verification tests, but this test will pull data blocks from all restore points (full and incremental) required for the restore point in question to boot. (e.g., if you have Full backup + 4 incremental points (restore points 1-5). Restore point 5 depends on data in points 3, 2, and 1 (full backup) to boot, so those backup files will be read from during SureBackup)

[redhorse]

Thank you for your answer!

Yes, the mount server is on the source site. There is a domain controller located on the remote site, that's the only server I could use as a mount server.

My question was related to the health check not to the Sure Backup job, but I guess it's the same and the whole chain is checked during the health check?

[david.domask]

Hi redhorse,

Ah yes, apologies - same concept applies, but I mis-wrote when I wrote SureBackup, but Health Check needs to mount data from the backup to check it, so same concept applies.

I would advise against doing this on a DC personally -- is there any way to get approval for a mount server? Even a virtualized one is fine, and the system requirements are fairly petite.

[redhorse]

Thanks for the explanation.

I feel the DC is the only option, which ports have to be open between the mount server and the hatdened repository to perform the health check via mount server? The hardened repository is currently isolated as far as possible.

[david.domask]

The ports required are listed on our User Guide page here. If the DC is the only option, give it a shot but please be aware it's against best practices.

[redhorse]

I've opened a case (#07692495) because the installation of the mount server on the DC didn't help. The first reply is that the mount server on the remote server (located next to the hardened repository) doesn't help to improve the speed of the health check.

I also observed that we have just 1M network usage on the backup server during the health check (and hardly CPU & RAM usage) even without the usage of a mount server on the remote site, so what is limiting the performance of the health check?