Synthetic backup or a Active Full Backup

[Jack1874]

Hi Guys, we are using a HP StoreOnce to land our data.

Our Backup Mode is set to Incremental

Should we use a weekly synthetic backup or a Active Full Backup ?

I feel that synthetic is the best method for us...

Thoughts?

[kte]

active full if it is you're primary target

[Jack1874]

Thanks Kte.

Could you explain a little bit more ...

[foggy]

Synthetic activity implies heavy random I/O on the storage, which typically does not mean good performance on dedupe storage. We are adding HP StoreOnce Catalyst integration in v9, though, which is intended to change the picture.

[Jack1874]

Thanks guys.. I appreciate the feeback.

[pshute]

Just trying to get my head around some Veeam concepts. If an active full back uses heaps more network bandwidth, etc, than a synthetic full backup, why ever do them, apart from the first time?

[PTide]

Hi, your post has been merged to an existing topic.

In addition to the info above you should keep in mind that when doing the synthetic full, Veeam uses the chain of incremental to build the full backup. Imagine one of the restore points has corrupt data thus your synthetic full might be unusable. That is why you should consider doing a periodic Active Full.

Thank you.

[pshute]

OK, thanks for merging that, I realised later I'd posted it in the wrong forum too.

So you're saying it's unwise to rely on a chain of incrementals that's too long?

[PTide]

No problem!

So you're saying it's unwise to rely on a chain of incrementals that's too long?

Exactly. Imagine that you have a chain of 7 daily incrementals with Full in the very end. In order to restore to the 5th day you need to have your Full + 4 subequent incrementals healthy. If any increment in forever-forward chain gets corrupted then you'll not be able no restore anything that is beyond it: F - i - i - i - X - i - i - i. The longer your chain exists without intermediate Fulls the more probability of loosing a big chunk of your chain. So it's better to perform periodic Fulls.

[larry]

If your backing up with CBT disabled will it find the bad X during a subsequent incremental?

[PTide]

Hi Larry,

If your backing up with CBT disabled will it find the bad X during a subsequent incremental?

Helpcenter:

During backup, Veeam Backup & Replication <...>scans through the VM image and calculates a checksum for every data block.
<...>
When incremental backup is run, Veeam Backup & Replication opens all backup files in the chain of previous full and incremental backups, reads metadata from these files and compares it with checksums calculated for a VM in its current state. If a match is found (which means the block already exists in the backup), the corresponding block is filtered out.

This basically means that if any block in the backup file gets corrupted the next job run will not detect that the blocks are different since checksum is compared to checksum from meta.

So CBT is not about error-correction, it's about excluding redundant data from backup. You can validate your backups with Backup Validator - it basically recalculates checksums for all data in the backup chain and compares the results to the previous value.

Thank you.

[larry]

Thanks I didn't know about the tool.
Does it have an exit code when all was successful? Could run as post script if exit code. It would be nice to run once a week when I now run the full. I want to do a full only once a month ( weekly now ) but I need to verify nothing has happened to the data while it is in storage. Is this what happens when I check "Enable automatic backup integrity checks"?

Lots of these issue came up when my raid controller failed on my backup repository. Some data got corrupt. The tool would of allowed me to verify the remaining data. I had off site copies and tapes but it was still very time consuming because I didn't trust any thing in the repository and rebuilt it from the copies.

[PTide]

It would be nice to run once a week when I now run the full.<...>I need to verify nothing has happened to the data while it is in storage.

Please check this thread.

Is this what happens when I check "Enable automatic backup integrity checks"?

No. Integrity check validates only metadata.

Also there is an option to use SureBackup job.

[larry]

thanks, will change the ps and run only on Saturday as post task, send email report on error, just what I wanted. On course if I trip a full backup on error will be better but will need to think more.