First Health Check on Windows Cluster not on Schedule

[AndrewTobin]

Hey all,

We have been running Veeam backup on our SQL Cluster for the last 3 nights - the first night it performed a full backup, the second night it performed a synthetic diff backup (as I understand it), and the third night it hasn't run, as the cluster host from the second night is still running a Health Check. For the settings of the job, I had set up the Health Check to run on the Last Friday of the month. That isn't the first Sunday morning (2:30am), so I am unsure why it is running now.

Does it run after every synthetic diff backup?

It appears that the files are locked during a health check and we can't restore during this (I was trying to restore a database to another server), is this true?

We may have to stop performing health checks and go with full backups weekly, if this is the case.

[HannesK]

Hello,

Does it run after every synthetic diff backup?

no. health checks only run on the configured days (as long as it takes).

If you are a SQL admin where differential backups are something normal, then there are a some differences to consider when talking about image-based (Veeam) backups. Veeam does incremental (and full if configured) backups. Every incremental backup is a logical full backup. The concept of differential backups does not exist because there are rare use-cases for it and differential backups add complexity.

There is a KB article with animations which explains the different backup modes

Question 1: are we talking about VM-based backup (VMware / Hyper-V) or agent-based backup with Veeam Agent for Windows?
Question 2: what kind of backup storage are you using that it takes days to finish a health check?
Question 3: which version of Backup & Replication are you using? I'm asking because health check speed was improved in 11a

Best regards,
Hannes

[AndrewTobin]

Hey Hannes,

Thanks for that, I meant to post some screenshots, but they weren't coming thru to the forum (showed as broken images), so I will try again.

I had a word with our infrastructure people and now understand the incrementals better - we're now performing a weekly full, and nightly incrementals, and no longer doing the health check, because we're getting ~14tb on the weekly full.

1. Agent based on a raw Windows Cluster.
2. NAS, with I believe 7200rpm disks, the full backup took around 10 hours, so it wasn't too bad.
3. 11.0.0.837

Settings Image
Running health check image

[HannesK]

Hello,

and no longer doing the health check, because we're getting ~14tb on the weekly full.

yes, with weekly active full, heath checks are unnecessary from my point of view

3. that's too old for the health check improvements

Best regards,
Hannes

[AndrewTobin]

Thanks Hannes,

That's what we're going to go with. I'm still a bit uncertain about why the health check started off schedule, but the theory on our end was that it started because it was the first incremental after the scheduled health check trigger, and so it decided to perform a health check ASAP, even though the job hadn't been set up until after that schedule trigger had passed, instead of waiting for the next "Last Friday of the month"