Hi Yuriy,

On Fri, Apr 21, 2017 at 12:59 PM, Yuriy Zveryanskyy <
[email protected]> wrote:

> Hi.
>
> After "ipminative" driver has been removed (I think it was right
> decision), we support IPMI in ironic only via drivers which use
> "ipmitool" utility.
> This utility is mostly good, but main problem is that running by
> ironic subprocess can be stalled on buggy/broken BMCs.
>

Here is one example of such issue -
https://bugs.launchpad.net/ironic/+bug/1683902, and a bit of comments about
the root cause in the eavesdrop
<http://eavesdrop.openstack.org/irclogs/%23openstack-ironic/%23openstack-ironic.2017-04-19.log.html#t2017-04-19T15:39:14>
.


> This causes situations like stop executing of sync power state
> periodic task without any logging, reduce free green threads
> number in the conductor service pool etc.
> Administrators often have only one version of ipmitool in
> repository and should build new version from source for
> bug fixing.
> We can implement custom executor for ipmitool with timeout
> for process, but this adds more complexity to IPMI drivers,
> or maybe use another solution? Maybe we should have pure
> Python well tested IPMI library optimized for ironic (like sushy
> for RedFish)?
>

Or maybe work on improving pyghmi, and reintroduce the ipminative driver.


>
> Yuriy Zveryanskyy
>
> __________________________________________________________________________
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: [email protected]?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: [email protected]?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to