Regarding Backup of Jira and all other Suites

[roopa shinde]

Hi,

I am new to Jira so have few queries regarding its backup.

Do i need to stop jira services when i take its backup? i have configured postgresql as the database.

We have confluence,crowd,fisheye,jira configured and I would like to backup this data using Symantec Backup Exec.

right now i have configured backup but the backup runs without stopping any of these services.

Can anyone please help me regarding how can i backup the atlassian products

[Andrzej Pasterczyk]

Hi,

See here http://confluence.atlassian.com/display/JIRA/Backing+Up+Data

I guess it depends on the tool you're using e.g. pg_dump runs in a serialized transaction, so it will always take a consistent backup (even if there are update/insert statements in same time).

[Andrzej Pasterczyk]

Then as I've mentioned pg_dump will take consistent backup, no need to stop any service. Of course depending on your hardware/software configuration and database size it may degrade performance during backup but should be relatively fast and if you do it at night time (or whatever time your server is not under heavy load) there will be no impact.

I would stick to hot backup with pg_dump, if for some reason you need to make a cold backup and take the database offline then obviously you should stop applications that access it.

[roopa shinde]

Hi,

yes i take backup of postgresql using pgdump and it is added to windows scheduler.

my concern is more on jira/confluence/crowd cold backup as now i am taking hot backups

[roopa shinde]

Hi,

Thank you so much for answering my queries. But there are attachments and other files in the data directory of jira home and other product home folders apart from postgresql database

I am concerned about the home folder backups of all these atlassian products. my query is if i take hot backup of the home directories, is it consistent when i restore it during worst condition like server breakdown?

Thanks again

[Andrzej Pasterczyk]

I'm not an expert but if you take the DB backup first and then the attachment directories the worst case scenario is that:

- you'll backup some files that were added after you've backed up DB but before you've copied the directory (so no harm just files that are not conntected to any item in DB backup)

- you'll have some files pointed out in DB that were removed from the filesystem and not in your backup (so you won't be able to view/download them even though Jira/Confluence will say they're there)... not too much harm I would say (since they were removed on puprose)

- don't know about other products like Crowd

It's basically a question of your expectations. If that's a system that needs to be available 24x7 you don't get too much choice (can't stop services). Otherwise you should prepare some nice scripts that will make sure it's up and running after backup and will notify you if it's not.