Running a teuthology workload depending on the distribution

[Loic Dachary]

Hi Zack,

The isa erasure code plugin is only built on some architectures. The workload described at

https://github.com/ceph/ceph-qa-suite/blob/master/erasure-code/ec-rados-plugin%3Disa-k%3D2-m%3D1.yaml

can therefore only be run on those architectures. How would you recommend that it is integrated in the upgrade suites ? Adding it to

https://github.com/ceph/ceph-qa-suite/tree/master/suites/upgrade/firefly-x/stress-split/9-workload

for instance, at the same level as

https://github.com/ceph/ceph-qa-suite/blob/master/suites/upgrade/firefly-x/stress-split/9-workload/ec-rados-plugin%3Djerasure-k%3D3-m%3D1.yaml

will not work because it will fail when running on precise (for instance).

Cheers

[Zack Cerza]

Hi Loic,

Sage recently filed these, which seem like they might help with what
you're asking.

http://tracker.ceph.com/issues/9256
http://tracker.ceph.com/issues/9255

I'll be getting to work on them relatively soon, just not quite yet.
In the meantime, would it be feasible to, say, make the relevant task
in the workload check for the presence of the plugin, and immediately
fail if it's not found? That way we can waive those failures, and have
them not use many more resources than necessary.

Zack

Post by Loic Dachary
Hi Zack,
The isa erasure code plugin is only built on some architectures. The =

workload described at

Post by Loic Dachary
https://github.com/ceph/ceph-qa-suite/blob/master/erasure-code/ec-rad=

os-plugin%3Disa-k%3D2-m%3D1.yaml

Post by Loic Dachary
can therefore only be run on those architectures. How would you recom=

mend that it is integrated in the upgrade suites ? Adding it to

Post by Loic Dachary
https://github.com/ceph/ceph-qa-suite/tree/master/suites/upgrade/fire=

fly-x/stress-split/9-workload

Post by Loic Dachary
for instance, at the same level as
https://github.com/ceph/ceph-qa-suite/blob/master/suites/upgrade/fire=

fly-x/stress-split/9-workload/ec-rados-plugin%3Djerasure-k%3D3-m%3D1.ya=
ml

Post by Loic Dachary
will not work because it will fail when running on precise (for insta=

nce).

Post by Loic Dachary
Cheers
--
Lo=C3=AFc Dachary, Artisan Logiciel Libre

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" i=
n
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

[Loic Dachary]

Hi Zack,

It's great that it's scheduled for this sprint, if I'm reading http://tracker.ceph.com/versions/378 correctly ;-) In the meantime it is not a blocker, I can work around it.

Thanks !

Post by Zack Cerza
Hi Loic,
Sage recently filed these, which seem like they might help with what
you're asking.
http://tracker.ceph.com/issues/9256
http://tracker.ceph.com/issues/9255
I'll be getting to work on them relatively soon, just not quite yet.
In the meantime, would it be feasible to, say, make the relevant task
in the workload check for the presence of the plugin, and immediately
fail if it's not found? That way we can waive those failures, and have
them not use many more resources than necessary.
Zack

Post by Loic Dachary
Hi Zack,
The isa erasure code plugin is only built on some architectures. The workload described at
https://github.com/ceph/ceph-qa-suite/blob/master/erasure-code/ec-rados-plugin%3Disa-k%3D2-m%3D1.yaml
can therefore only be run on those architectures. How would you recommend that it is integrated in the upgrade suites ? Adding it to
https://github.com/ceph/ceph-qa-suite/tree/master/suites/upgrade/firefly-x/stress-split/9-workload
for instance, at the same level as
https://github.com/ceph/ceph-qa-suite/blob/master/suites/upgrade/firefly-x/stress-split/9-workload/ec-rados-plugin%3Djerasure-k%3D3-m%3D1.yaml
will not work because it will fail when running on precise (for instance).
Cheers
--
Loïc Dachary, Artisan Logiciel Libre