[Openstack-operators] memcached redundancy
clint at fewbar.com
Thu Aug 21 19:08:47 UTC 2014
Excerpts from Joe Topjian's message of 2014-08-21 11:57:44 -0700:
> Hi Clint,
> Thank you for your input.
> If I understand you correctly, the core cause seems to be internal to
> OpenStack? If that's true, I will create a bug report about this. I'm
> guessing Oslo would be the correct project to file the bug?
Oslo is where the "memorycache" that nova uses resides. I imagine
others use it too. It is entirely possible, though, that each usage of
the module itself could require refactoring so that each process keeps a
single cache connection object alive rather than creating one per request,
and so that each usage of it can make use of dead_retry.
More information about the OpenStack-operators