Health check fundamentals?

[nokogerra]

Good day.
I`m confused about a health check process details so may be someone can help me to understand this process.
1. I really can`t understand the point of this operation. CRC and hash values is calculated twice for the same restore point and then compared, am I right?
2.

The health check is performed at the beginning of the synchronization interval.

What is the synchronization interval?
3.

When Veeam Backup & Replication saves a restore point to the backup repository, it calculates CRC values for backup metadata and hash values for data blocks of VM disk in the backup file and saves these values in the metadata of the backup file, together with VM data.

Does Veeam calculate that CRC values and hash values at the end of each backup job independently of health check configuration or it calculates that values only if health check is enabled?

[foggy]

1. First time it is calculated during the time the backup file is created and stored on disk along with the backup itself. So CRC data is kept for data blocks in every backup file in the chain. When the health check is started, CRC for all the blocks comprising the most recent most restore point (that are scattered along multiple files) are calculated again and compared with those already stored on disk.
2. Synchronization interval is the time interval you specify for the backup copy job to sync data.
3. Independently at each job cycle.

[nokogerra]

Thanks for your reply.
If "synchronization interval" is about a copy job, then it is not applicable for my situation. I do not use copy jobs.
I`ve noticed that the health check is performed after the backup job is completed, e.g. http://hkar.ru/KkQL. I thought that the health check should be performed before the backup job is completed (it says here https://habrahabr.ru/company/veeam/blog/276815/ that health check really should be performed before the backup job). I mean, whats the point to check the health of the restore point that was made just now?

And what if the health check is scheduled on the day which the backup job is not performed? Will the health check be performed all the same?

update: also I would like to know how comes that health check takes pretty long time to perform for significant restore points? If CRC and hash values are stored after each backup job, I guess it takes a small time period (I didn`t notice any considerable time period at the end of the backup job). So why the health check itself takes so much time if the calculation of CRC and hash actually takes insignificant time period? Is it about a comparison period?

[foggy]

I`ve noticed that the health check is performed after the backup job is completed, e.g. http://hkar.ru/KkQL. I thought that the health check should be performed before the backup job is completed (it says here https://habrahabr.ru/company/veeam/blog/276815/ that health check really should be performed before the backup job).

It's different for the backup and backup copy jobs. For the backup copy jobs, health check is performed at the beginning of the synchronization interval, while for the backup job - at the end of the job session.

I mean, whats the point to check the health of the restore point that was made just now?

Blocks comprising the restore point might be located in different backup files, so health check validates all of them (not just the ones that were written during the last session).

update: also I would like to know how comes that health check takes pretty long time to perform for significant restore points? If CRC and hash values are stored after each backup job, I guess it takes a small time period (I didn`t notice any considerable time period at the end of the backup job). So why the health check itself takes so much time if the calculation of CRC and hash actually takes insignificant time period? Is it about a comparison period?

CRC values are calculated for each block on-the-fly during the backup, at the time it is being processed by the job, so doesn't take much time. While during health check each block has to be read to calculate it's CRC, so this process requires reading the entire restore point, which is time consuming.

[foggy]

And what if the health check is scheduled on the day which the backup job is not performed? Will the health check be performed all the same?

In this case, health check will be performed after the next job session.

[nokogerra]

Thank you for the comprehensive reply.

In this case, health check will be performed after the next job session.

I had doubt about it because faced similar problems with the synthetic full bakup jobs recently. E.g. the synthetic full jobs was scheduled at saturday, however during wrong scheduling options (chaining) some jobs was not performed at this day and as a consequence the synthetic full backups were not performed at all.

[foggy]

But you were asking about the health check behavior. Synthetic fulls are different.

[nokogerra]

Well if it is so, I guess I got all the info I need, thanks a lot for your help.