JIRA FAIL OVER implementaion

[Nagaraja]

HI

I am trying to implement JIRA failover in linux machine. I need help in file replication.

I have extracted JIRA in my Home directory of linux. Please let me know what are the files need to be replicated to support failover.

We have SAN file over approch and directory is in different.

How to link those directories?

[AhmadDanial]

Hey there, Nagaraja.

Perhaps the following documentation can shed extra light in addition to the explanation provided by Nic earlier on:

https://confluence.atlassian.com/display/ATLAS/Failover+for+JIRA

Hope it helps!

Warm regards,

Danial

[Bryan Karsh]

Grrr. This page no longer exists. Every approach I take leads me here: https://confluence.atlassian.com/display/ENTERPRISE/Failover+for+JIRA+Data+Center -- which is not helpful for those of use who employ cold failover. Also... since Data Center solution doesn't support clustering between different geographical locations, there are some of us who'd prefer a cold failover type configuration.

[Nic Brough -Adaptavist-]

Jira has several parts, so it depends on how you've installed it and where.

The application is pretty much fixed in terms of disk access, you can replicate the application directories and forget them. You should have installed that in a separate area to the "jira home" directory.

You must replicate the content of the database, wherever that is

Now, under "jira home", there's several sections

You must replicate the content of the attachments directory if you are using them - you can see/set the path for those in the admin screens, but it will be under the "jira home"

You should replicate the plugins directory, or you'll need to reinstall them after failover

There are temporary files that you don't need in other directories, the application will rebuild them, so don't worry about them

Finally, there's the index. Technically, you should replicate it if you want a near-instant failover. In real life, putting it on a SAN can slow the access to the point of unusability, so it might be impossible. You need to test that before you replicate it.

[Nagaraja]

Thanks Nic.

If i ask specifically .

I have installed JIRA in /home/svnuser/Jira/Jira Home directory

The share file system is available is /GSSVN/ directory

so if failover occure /GSSVN file system is shared between two server and all the files will be moved to other linux server

So as per your answer i need to install jira in /GSSVN

or anyway i can replicate the files using /GSSVN folder

[Nic Brough -Adaptavist-]

Yes, but you need to sub-divide that share into at the very least Application and Data.

And, you really really need to test putting the index on it - there is a VERY good chance a shared/replicated index will NOT work.