Why does snapshot overflow occur?

[shahn]

Hi.

I try backup on standalone Linux.

But failed snapshot overflow occured.

Why does snapshot overflow occur?

I know that correcting stretch value will solve it
but, i want to know why.

[david.domask]

Hi shahn,

Snapshot overflow typically occurs when there is sizeable change rate for the backup and the default settings for the snapshot aren't enough. There are several retries built in with larger snapshots, but if the retries aren't successful, then the changes can help. If it's happening frequently, it's best to open a support case to check the behavior. (Please remember to include logs from the Linux Agent job/source machine for Support to review)

But in general correcting the value is fine as I understand it.

[ybarrap2003]

Here is another similar issue that occured with Rocky Linux. Are you using a support Linux OS? - veeam-agents-for-linux-mac-aix-solaris- ... 87593.html

[aj_potc]

Unfortunately that thread was never updated with a resolution, so it's hard to know if these cases are connected with the Linux distro.

In my own case, starting with Veeam Agent for Linux 6 I got the snapshot overflow error on every backup attempt on a certain Web server. This is despite the fact the system should be relatively "quiet" from the standpoint of data changes during the backup window. (Apparently Veeam disagrees.)

I got it working by setting some fixed min and max snapshot sizes (5 GB and 50 GB, respectively). My choices were pretty much arbitrary.

Ultimately, I set the following values in /etc/veeam/veeam.ini:

Code: Select all

maxSize = 53687091200
type = common
location = /tmp
minSize = 5073741824

[PTide]

Hi @aj_potc

Have you seen this post?

Snapshot overflows are distro-agnostic - not a single case of overflow over the course of ~8 years has been found to be connected to any distro in particular.
There two algorithms for snapshot allocation - 'stretch' (default) and 'common'.

Stretch snapshot grows as needed.
Common is pre-allocated at the start of the backup job and cannot grow.

Changing the stretch portion size parameter ('portionSize') can help in cases when the snapshot segments get filled faster than VAL manages to allocate another segment.

By using the values above you have restricted the agent to use only 'common' algorithm.

In addition to that, your setting impose limitations on the snapshot size (upper and lower) and also bind the snapshot to the specific place on a filesystem (/tmp).

Thanks!

[aj_potc]

Hi Pavel,

Yes, I'm aware of the effects of my changes and did so only after consulting posts here and Veeam documentation.

The "stretch" snapshot method didn't work for one of my systems, so I suggested how a switch from "stretch" to "common" alleviated the problem for me. Maybe I could have gotten "stretch" to work after enough tries, but sometimes you don't have time to play around with these parameters on a production system, and you just need to get things working. In that case I'm fine with using the "common" method to reserve some fixed space for a snapshot.

[shahn]

Thank you for all coments.