Using File System snapshots to backup up Atlassian Stash

[Geoff Bush]

We have been testing the following procedure, which we do use to backup other applications, to do an overnight backup of Atlassian Stash:

We have configured both STASH_HOME and the metadata database (PostgresSQL) to store their data on the same Network Appliance (NetApp) data filer volume. We then do the following:

01:55Z Stop Atlassian Stash

01:57Z Stop PostgreSQL

02:00Z Snapshot of the NetApp data volume taken (basically an imutable snapshot of the filesystem at that moment in time)

02:03Z Start PostgreSQL

02:05:Z Start Atlassian Stash

Note: We do understand that this is a full copy roll back backup and we would only use this for Disaster Recovery puposes.

What we want to know is, do Atlassian recognise/approve of this form of backup?

[Roger Barnes · Accepted solution]

From a technical viewpoint, for DR purposes as you say, this sounds like a good approach to take. Furthermore, our documentation on backups does cover this approach to cold (offline) backups, so it is certainly recognised in that respect. As long as you're backing up (and restoring) in sync as described, and are comfortable that a NetApp snapshot suits your backup needs for DR purposes, then I'd say you are in good stead.