MYSQL LOCK every one hour at same time

[Sarath Kumar K]

Hi Guys,

I am using mitaka (installed using RDO) 1 Controller and 4 compute nodes. And i am running almost 20 instances.

Every one hour mysql lock is happening. attaching keystone.log.

This is happening because number of mysql queries/threads are increasing at this particular time.

Thanks in advance

[Steve Martinelli (stevemar)]

The reason it's once per hour is because of the cron job likely setup with your distro. What happens when you perform the keystone-manage token_flush command from a shell?

[Sarath Kumar K (kumar-sarath588)]

Hi Steve,

Can you suggest, how i could reduce the lock as nagios monitoring (which
monitores all the services of openstack queries few things like number of
instances, glance image list etc) shows critical every one hour as the
queries fail.

Thanks.

On Thu, Aug 4, 2016 at 1:18 AM, Steve Martinelli <<email address hidden>
> wrote:

> The reason it's once per hour is because of the cron job likely setup
> with your distro. What happens when you perform the keystone-manage
> token_flush command from a shell?
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1609524
>
> Title:
> MYSQL LOCK every one hour at same time
>
> Status in OpenStack Identity (keystone):
> New
>
> Bug description:
> Hi Guys,
>
> I am using mitaka (installed using RDO) 1 Controller and 4 compute
> nodes. And i am running almost 20 instances.
>
> Every one hour mysql lock is happening. attaching keystone.log.
>
> This is happening because number of mysql queries/threads are
> increasing at this particular time.
>
> Thanks in advance
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/keystone/+bug/1609524/+subscriptions
>

[Steve Martinelli (stevemar)]

Long term, as your deployment grows you probably want to switch to using fernet tokens as they are not persisted, you won't need to run that cron job at all. But doing this increases the complexity of running keystone, so let's put that aside for now.

Have you tried simply reducing the frequency of the cron job? How many users are able to access the cloud? Running once per hour for 20 instances seems unnecessarily aggressive.

[Launchpad Janitor (janitor)]

[Expired for OpenStack Identity (keystone) because there has been no activity for 60 days.]