[Openstack-operators] Database cleanup policy

Morgan Fainberg morgan.fainberg at gmail.com
Thu Oct 30 20:47:08 UTC 2014

Hi Abel,

For Keystone we already have a way to prune out expired records: keystone-manage token_flush

This can be run via cron (recommended). The reason for the side band tool is that keystone does not have an internal scheduler for periodic tasks (not a common use keystone needs to do across all the functionality) 

If you have a large number of tokens and use MYSQL, we have logic to help limit he impact to the backend by doing batched flushes.

I am not sure what the requirements for holding on to data (e.g. Nova instances) once they've been deleted, but I think it is definitely worth setting some clear guidelines on this for each service so it can be followed / implemented as a built in function.


Sent via mobile

> On Oct 30, 2014, at 13:20, Abel Lopez <alopgeek at gmail.com> wrote:
> We just had this question come up regarding the labs, but it applies to production as well.
> I'm thinking that we need to implement some sort of periodic database pruning. Perhaps like every two months or so, go through all databases, all tables, and do like
> delete from FOO where deleted=1 and deleted_at < date_sub(now(), interval 2 month)
> Just as an example.
> Does anyone have see any issues with purging deleted=1 data?
> I've seen this be very helpful for things like Keystone tokens, etc.
> _______________________________________________
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20141030/76c25191/attachment.html>

More information about the OpenStack-operators mailing list