Recently i came across an issue with the Linux nosnap Agent running on an IBM Power (ppc) architecture.
So since those systems are not so common as x86 there is not a whole lot Information out there, why i thought this could be interesting for some.
So first of all: The rollout of the Agent is not your typical “Create Protection Group-->Push Agent” but you have to do it through the “Predefined Agent” method:
https://helpcenter.veeam.com/docs/backup/agents/protection_group_packages.html?ver=120
The customers systems hosted Databases (MAXDB) which we backed up in a seperate job with a pre script. This Script dumped the DB and Logs and after that we backed up the Volume.
Which means we wanted to create a second job which backed up only the remaining volumes, so we chose “Volume Level Backup”.
The Issue was that there is a limitation in the documentation:
https://helpcenter.veeam.com/docs/agentforlinux/userguide/system_requirements_nosnap.html?ver=60
- The amount of space required for LVM snapshots largely depends on the IO intensity. Generally, from 10% to 20% of the system’s occupied space should be enough for storing an LVM snapshot.
But the Customer had a lot of volumes which did not have enough space and wasnt able to enlarge them.
When we chose “Entire Computer” we did not run into the problem, but that meant that we back up Data twice since we backed up the DB already seperately.
The Solution was in this KB the 3rd Bullet Point:
We added the DB volumes to the ignoreDevices, restarted the veeamservice as described, after that the Entire Computer Backup job ran without Issue and ignored the Volumes.
So a corner case for sure, but maybe someone will find this useful sometime.
Cheers
SpeedyJ
