I am interested in understand how the community is backing up their meter.
Personally I am going with the approach of VM snapshot based backup, and also dumping the DB in the file system via cron.daily bash script.
The user guide shows how to dump the DB:
To back up the database
Run this command as the PostgreSQL user:
pg_dump usage_meter_db | gzip > /tmp/usage_meter_db.pgd.gz
This will auto-gzip the database into a single file that can be viewed.
I was going to extend this to running a backup agent inside the appliance (which would run the script pre-backup). We have an Avamar backup system which supports SUSE. The only problem is missing dependancies.
libcap.so.2()(64bit) is needed by AvamarClient-6.0.101-66.x86_64
Does anyone else run backup agents inside the appliance?
Can anyone specifically help about the missing dependancy?
Any feedback welcome.
A database dump is a great way to back up Usage Meter. I would strongly recommend against using VM snapshots as backups for Usage Meter or any VM for that matter. Having a long snapshot chain attached to a VM can lead to performance problems, as well as increase your risk for data loss and potentially even corruption. Snapshots are really intended for use when upgrading an application, applying patches, or performing operations in which you may need to quickly revert back to a state prior to the actions you're performing. They are also, often used by backup solutions to quiesce the VMDK while taking a backup of the VM since all changes are written to a separate file and allows the backup software to backup the base disk. I will keep an eye out for viable backup solutions for Usage Meter, but I have yet to see any personally. I have included a link below which discusses snapshots in more detail. Forgive me if I am misunderstanding what you mean by "VM snapshot based backup."
Technical Support Engineer | VCP4
VMware Global Support Services