[ubuntu] Veeam agent for linux and high data churn rate during snapshot creation (Solved)

[pete]

I'm testing moving our monitoring server backups from Bacula to Veeam, but the high churn rate (rrd files) was causing the snapshot to grow too large and the backup fails with snapshot overflow messages.

 

The fix (for me, YMMV) was to open /etc/veeam/veeam.ini and edit the following line:

 

portionSize=

 

...by doubling the bytes value (as a starting point). The value is calculated at install time, but if your server has a lot of data changing (rrd files were the culprit for me) it may be too low.

 

Found here: https://forums.veeam.com/veeam-agent-for-linux-f41/snapshot-overflow-t36609-90.html after some helpful person reposted the guidance they'd received from support. Less helpful were the Veeam employees who took the "ooh, it's really tricky, contact support" stance then admitted it's usually not that tricky.

[PTide]

Hi Pete

 

Right, it's not tricky at all. Honestly speaking, I don't remember anyone ever saying on Veeam forums that it was tricky. But, exactly as you've said:

 

1) YMMV

2) Statistics. Without people contacting support team (which is available for free edtition as well, btw) Veeam would have almost no idea of what adjustment should be introduced in future releases, and what are typical setups and sizes.

 

Anyway, I'm glad that you've nailed it!

 

Thanks

[pete]

Hi Pete

 

Right, it's not tricky at all. Honestly speaking, I don't remember anyone ever saying on Veeam forums that it was tricky. But, exactly as you've said:

 

1) YMMV

2) Statistics. Without people contacting support team (which is available for free edtition as well, btw) Veeam would have almost no idea of what adjustment should be introduced in future releases, and what are typical setups and sizes.

 

Anyway, I'm glad that you've nailed it!

 

Thanks

 

Hi PTide,

 

This was documented as a known issue in the release notes for 2.0, so I assumed there'd be a KB article expanding on that. Instead I found a 7-page thread with people being asked to open support tickets. Other vendors I deal with would have issued general guidance (try X or Y, see this KB article, open a ticket if that still doesn't help) before the thread got that long if they knew what the fix is most of the time.