Jira / Crowd cache question (re: contigency for crowd failure)

[Bryan Karsh]

Hi guys, Here is a question. I am running Jira 5.2.6 / Crowd 2.6.0 I was wondering if it is possible to increase the caching on the Jira end -- for example, instead of 3600 seconds, to have it be like several hours, or even a day. I am thinking in terms of what would happen if Crowd failed -- and we had to to failover/rebuild etc... it would be nice if we didn't have to worry about authenticating to Jira while we fixed Crowd. Anyone ever do something like this? Any cons? (other than crowd changes taking longer to propagate to jira?) Alternative strategies appreciated as well.

[Harry Chan · Accepted solution]

Hi, I would have a cold failover of Crowd ready. Set a low TTL on the DNS. You can literally failover in 5-10mins without any issues that way. Would solve most of the problems.

You can set the cache to longer. By default it's 1 hour isn't it? Unless you've changed it.

The downside of a longer cache is taking longer to propagate though you can manually refresh if you know there's a change you need to be picked up.