Status of luminous v12.2.1 QE validation

Status of luminous v12.2.1 QE validation

[Yuri Weinstein]

Details of this release summarized here
http://tracker.ceph.com/issues/21296#note-18

The following suites included:

rados
rgw
rbd
fs
krbd
kceph
knfs
rest
hadoop - EXCLUDED
samba - EXCLUDED
ceph-deploy
ceph-disk
upgrade/client-upgrade
upgrade/jewel-x (luminous)
upgrade/kraken-x (luminous)
upgrade/luminous-x (master)
powercycle
ceph-ansible
(please speak up if something is missing)

Issues:

rados.rbd
http://tracker.ceph.com/issues/21428 (assume is a `must` to be backported)

fs
http://tracker.ceph.com/issues/21381 - known
http://tracker.ceph.com/issues/20910 - known
http://tracker.ceph.com/issues/20736 - known

kcephfs - needs analysis - Patrick, Zheng FYI

knfs
http://tracker.ceph.com/issues/21424

rest
http://tracker.ceph.com/issues/21425

ceph-ansible related - per Vasu ceph-asible pass on master, PRs are to
be backported:
https://github.com/ceph/ceph/pull/17757
https://github.com/ceph/ceph-ansible/issues/1834

RHEL 7.4 related PRs need to backported:
https://github.com/ceph/ceph/pull/17776
https://github.com/ceph/ceph/pull/17796
https://github.com/ceph/ceph/pull/17777

Not all tests were completed yet.

I suggest to consider the list for fixing/backporting and then rerun
all suites on a new sha1.

Thx
YuriW

Re: Status of luminous v12.2.1 QE validation

[Nathan Cutler]

Hi Yuri:

> RHEL 7.4 related PRs need to backported:
> https://github.com/ceph/ceph/pull/17776

Luminous backport PR is open: https://github.com/ceph/ceph/pull/17779

> https://github.com/ceph/ceph/pull/17796

The PR you mentioned (17796) is the luminous backport PR for 
https://github.com/ceph/ceph/pull/17795 (the latter will presumably be 
merged into master in short order).

> https://github.com/ceph/ceph/pull/17777

Luminous backport PR is open: https://github.com/ceph/ceph/pull/17816

Hope this helps.

Nathan

Re: Status of luminous v12.2.1 QE validation

[Yuri Weinstein]

Sure does, thx Nathan :)

On Tue, Sep 19, 2017 at 1:19 PM, Nathan Cutler <ncutler@suse.cz> wrote:
> Hi Yuri:
>
>> RHEL 7.4 related PRs need to backported:
>> https://github.com/ceph/ceph/pull/17776
>
>
> Luminous backport PR is open: https://github.com/ceph/ceph/pull/17779
>
>> https://github.com/ceph/ceph/pull/17796
>
>
> The PR you mentioned (17796) is the luminous backport PR for
> https://github.com/ceph/ceph/pull/17795 (the latter will presumably be
> merged into master in short order).
>
>> https://github.com/ceph/ceph/pull/17777
>
>
> Luminous backport PR is open: https://github.com/ceph/ceph/pull/17816
>
> Hope this helps.
>
> Nathan

Re: Status of luminous v12.2.1 QE validation

[Abhishek]

On 2017-09-19 22:10, Yuri Weinstein wrote:
> Details of this release summarized here
> http://tracker.ceph.com/issues/21296#note-18
> 
> The following suites included:
> 
> rados
> rgw
> rbd
> fs
> krbd
> kceph
> knfs
> rest
> hadoop - EXCLUDED
> samba - EXCLUDED
> ceph-deploy
> ceph-disk
> upgrade/client-upgrade
> upgrade/jewel-x (luminous)
> upgrade/kraken-x (luminous)
> upgrade/luminous-x (master)
> powercycle
> ceph-ansible
> (please speak up if something is missing)
> 
> Issues:
> 
> rados.rbd
> http://tracker.ceph.com/issues/21428 (assume is a `must` to be 
> backported)
> 
> fs
> http://tracker.ceph.com/issues/21381 - known
> http://tracker.ceph.com/issues/20910 - known
> http://tracker.ceph.com/issues/20736 - known
> 
> kcephfs - needs analysis - Patrick, Zheng FYI
> knfs
> http://tracker.ceph.com/issues/21424
> 
> rest
> http://tracker.ceph.com/issues/21425
> 
> ceph-ansible related - per Vasu ceph-asible pass on master, PRs are to
> be backported:
> https://github.com/ceph/ceph/pull/17757
> https://github.com/ceph/ceph-ansible/issues/1834

This also probably can wait, but if we are going to do a rebuild we can 
add this to the queue
> RHEL 7.4 related PRs need to backported:
> https://github.com/ceph/ceph/pull/17776
> https://github.com/ceph/ceph/pull/17796
> https://github.com/ceph/ceph/pull/17777

For the RADOS PR, which I'll defer to Sage/Josh to take a call,
the RGW hadoop s3 suite doesn't probably warrant a rebuild + rerun on 
its own. (Yehuda/Matt thoughts?)

> Not all tests were completed yet.
> 
> I suggest to consider the list for fixing/backporting and then rerun
> all suites on a new sha1.
> 
> Thx
> YuriW

Re: Status of luminous v12.2.1 QE validation

[Patrick Donnelly]

On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein <yweinste@redhat.com> wrote:
> kcephfs - needs analysis - Patrick, Zheng FYI

http://tracker.ceph.com/issues/21463
http://tracker.ceph.com/issues/21462
http://tracker.ceph.com/issues/21466
http://tracker.ceph.com/issues/21467
http://tracker.ceph.com/issues/21468

Last 3 look like blockers. Zheng will have more input.

-- 
Patrick Donnelly

Re: Status of luminous v12.2.1 QE validation

[Josh Durgin]

On 09/19/2017 01:45 PM, Abhishek wrote:
> On 2017-09-19 22:10, Yuri Weinstein wrote:
>> Details of this release summarized here
>> http://tracker.ceph.com/issues/21296#note-18
>>
>> The following suites included:
>>
>> rados
>> rgw
>> rbd
>> fs
>> krbd
>> kceph
>> knfs
>> rest
>> hadoop - EXCLUDED
>> samba - EXCLUDED
>> ceph-deploy
>> ceph-disk
>> upgrade/client-upgrade
>> upgrade/jewel-x (luminous)
>> upgrade/kraken-x (luminous)
>> upgrade/luminous-x (master)
>> powercycle
>> ceph-ansible
>> (please speak up if something is missing)
>>
>> Issues:
>>
>> rados.rbd
>> http://tracker.ceph.com/issues/21428 (assume is a `must` to be 
>> backported)
>>
>> fs
>> http://tracker.ceph.com/issues/21381 - known
>> http://tracker.ceph.com/issues/20910 - known
>> http://tracker.ceph.com/issues/20736 - known
>>
>> kcephfs - needs analysis - Patrick, Zheng FYI
>> knfs
>> http://tracker.ceph.com/issues/21424
>>
>> rest
>> http://tracker.ceph.com/issues/21425
>>
>> ceph-ansible related - per Vasu ceph-asible pass on master, PRs are to
>> be backported:
>> https://github.com/ceph/ceph/pull/17757
>> https://github.com/ceph/ceph-ansible/issues/1834
> 
> This also probably can wait, but if we are going to do a rebuild we can 
> add this to the queue
>> RHEL 7.4 related PRs need to backported:
>> https://github.com/ceph/ceph/pull/17776
>> https://github.com/ceph/ceph/pull/17796
>> https://github.com/ceph/ceph/pull/17777
> 
> For the RADOS PR, which I'll defer to Sage/Josh to take a call,

This one got through testing today, with runs almost entirely devoid of 
noise. If there are odd hangs in other suites that may have been caused
by this, you may want to try re-running. rados and rbd suites look good
at this point though.

Re: Status of luminous v12.2.1 QE validation

[Abhishek]

On 2017-09-19 23:11, Patrick Donnelly wrote:
> On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein <yweinste@redhat.com> 
> wrote:
>> kcephfs - needs analysis - Patrick, Zheng FYI
> 
> http://tracker.ceph.com/issues/21463
> http://tracker.ceph.com/issues/21462
> http://tracker.ceph.com/issues/21466
> http://tracker.ceph.com/issues/21467
> http://tracker.ceph.com/issues/21468
> 
> Last 3 look like blockers. Zheng will have more input.

Here are the results of a rerun with -k testing,
http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/
& details http://tracker.ceph.com/issues/21296#note-20
There are a few failures which look environmental but a few look like 
they are related to the changeset (the ones with cache)

Best,
Abhishek

Re: Status of luminous v12.2.1 QE validation

[Abhishek]

On 2017-09-25 03:08, Yan, Zheng wrote:
> On 2017/9/24 22:03, Abhishek wrote:
>> On 2017-09-19 23:11, Patrick Donnelly wrote:
>>> On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein <yweinste@redhat.com> 
>>> wrote:
>>>> kcephfs - needs analysis - Patrick, Zheng FYI
>>> 
>>> http://tracker.ceph.com/issues/21463
>>> http://tracker.ceph.com/issues/21462
>>> http://tracker.ceph.com/issues/21466
>>> http://tracker.ceph.com/issues/21467
>>> http://tracker.ceph.com/issues/21468
>>> 
>>> Last 3 look like blockers. Zheng will have more input.
>> 
>> Here are the results of a rerun with -k testing,
>> http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/ 
>> & details http://tracker.ceph.com/issues/21296#note-20
>> There are a few failures which look environmental but a few look like 
>> they are related to the changeset (the ones with cache)
>> 
> "MDS cache is too large" issue should be fixed by
> https://github.com/ceph/ceph/pull/17922
> 
Ah alright this is a qa suite fix, we can run a the suite with a qa 
suite argument just to be sure.

Are there any other cephfs patches that needs to go in to the luminous 
brancha as such? I believe we would want the qe suites after the RADOS 
PR went in (https://github.com/ceph/ceph/pull/17796)

Sage, Josh, Yehuda, Patrick
Is there anything else that needs to go in before we start a second 
round?




> Regards
> Yan, Zheng
> 
>> Best,
>> Abhishek

Re: Status of luminous v12.2.1 QE validation

[Abhishek]

On 2017-09-25 11:43, Abhishek wrote:
> On 2017-09-25 03:08, Yan, Zheng wrote:
>> On 2017/9/24 22:03, Abhishek wrote:
>>> On 2017-09-19 23:11, Patrick Donnelly wrote:
>>>> On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein 
>>>> <yweinste@redhat.com> wrote:
>>>>> kcephfs - needs analysis - Patrick, Zheng FYI
>>>> 
>>>> http://tracker.ceph.com/issues/21463
>>>> http://tracker.ceph.com/issues/21462
>>>> http://tracker.ceph.com/issues/21466
>>>> http://tracker.ceph.com/issues/21467
>>>> http://tracker.ceph.com/issues/21468
>>>> 
>>>> Last 3 look like blockers. Zheng will have more input.
>>> 
>>> Here are the results of a rerun with -k testing,
>>> http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/ 
>>> & details http://tracker.ceph.com/issues/21296#note-20
>>> There are a few failures which look environmental but a few look like 
>>> they are related to the changeset (the ones with cache)
>>> 
>> "MDS cache is too large" issue should be fixed by
>> https://github.com/ceph/ceph/pull/17922
> Ah alright this is a qa suite fix, we can run a the suite with a qa
> suite argument just to be sure.

alternatively we can just merge this in luminous branch and this gets 
tested in the final qe run anyway, since we may have to do that since 
there are some rados and cephfs changes?
> 
  Are there any other cephfs patches that needs to go in to the luminous
> brancha as such? I believe we would want the qe suites after the RADOS
> PR went in (https://github.com/ceph/ceph/pull/17796)
> Sage, Josh, Yehuda, Patrick
> Is there anything else that needs to go in before we start a second 
> round?
> 
> 
> 
>> Regards
>> Yan, Zheng
>> 
>>> Best,
>>> Abhishek
> 
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" 
> in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Status of luminous v12.2.1 QE validation

[Sage Weil]

On Mon, 25 Sep 2017, Abhishek wrote:
> On 2017-09-25 11:43, Abhishek wrote:
> > On 2017-09-25 03:08, Yan, Zheng wrote:
> > > On 2017/9/24 22:03, Abhishek wrote:
> > > > On 2017-09-19 23:11, Patrick Donnelly wrote:
> > > > > On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein <yweinste@redhat.com>
> > > > > wrote:
> > > > > > kcephfs - needs analysis - Patrick, Zheng FYI
> > > > > 
> > > > > http://tracker.ceph.com/issues/21463
> > > > > http://tracker.ceph.com/issues/21462
> > > > > http://tracker.ceph.com/issues/21466
> > > > > http://tracker.ceph.com/issues/21467
> > > > > http://tracker.ceph.com/issues/21468
> > > > > 
> > > > > Last 3 look like blockers. Zheng will have more input.
> > > > 
> > > > Here are the results of a rerun with -k testing,
> > > > http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/
> > > > & details http://tracker.ceph.com/issues/21296#note-20
> > > > There are a few failures which look environmental but a few look like
> > > > they are related to the changeset (the ones with cache)
> > > > 
> > > "MDS cache is too large" issue should be fixed by
> > > https://github.com/ceph/ceph/pull/17922
> > Ah alright this is a qa suite fix, we can run a the suite with a qa
> > suite argument just to be sure.
> 
> alternatively we can just merge this in luminous branch and this gets tested
> in the final qe run anyway, since we may have to do that since there are some
> rados and cephfs changes?

Yes, let's do that.

I think all of the other blockers are resolved now?

sage


> > 
>  Are there any other cephfs patches that needs to go in to the luminous
> > brancha as such? I believe we would want the qe suites after the RADOS
> > PR went in (https://github.com/ceph/ceph/pull/17796)
> > Sage, Josh, Yehuda, Patrick
> > Is there anything else that needs to go in before we start a second round?
> > 
> > 
> > 
> > > Regards
> > > Yan, Zheng
> > > 
> > > > Best,
> > > > Abhishek
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 

Re: Status of luminous v12.2.1 QE validation

[Yuri Weinstein]

Do we do 2d round of QE validation ?

On Mon, Sep 25, 2017 at 10:28 AM, Sage Weil <sweil@redhat.com> wrote:
> On Mon, 25 Sep 2017, Abhishek wrote:
>> On 2017-09-25 11:43, Abhishek wrote:
>> > On 2017-09-25 03:08, Yan, Zheng wrote:
>> > > On 2017/9/24 22:03, Abhishek wrote:
>> > > > On 2017-09-19 23:11, Patrick Donnelly wrote:
>> > > > > On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein <yweinste@redhat.com>
>> > > > > wrote:
>> > > > > > kcephfs - needs analysis - Patrick, Zheng FYI
>> > > > >
>> > > > > http://tracker.ceph.com/issues/21463
>> > > > > http://tracker.ceph.com/issues/21462
>> > > > > http://tracker.ceph.com/issues/21466
>> > > > > http://tracker.ceph.com/issues/21467
>> > > > > http://tracker.ceph.com/issues/21468
>> > > > >
>> > > > > Last 3 look like blockers. Zheng will have more input.
>> > > >
>> > > > Here are the results of a rerun with -k testing,
>> > > > http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/
>> > > > & details http://tracker.ceph.com/issues/21296#note-20
>> > > > There are a few failures which look environmental but a few look like
>> > > > they are related to the changeset (the ones with cache)
>> > > >
>> > > "MDS cache is too large" issue should be fixed by
>> > > https://github.com/ceph/ceph/pull/17922
>> > Ah alright this is a qa suite fix, we can run a the suite with a qa
>> > suite argument just to be sure.
>>
>> alternatively we can just merge this in luminous branch and this gets tested
>> in the final qe run anyway, since we may have to do that since there are some
>> rados and cephfs changes?
>
> Yes, let's do that.
>
> I think all of the other blockers are resolved now?
>
> sage
>
>
>> >
>>  Are there any other cephfs patches that needs to go in to the luminous
>> > brancha as such? I believe we would want the qe suites after the RADOS
>> > PR went in (https://github.com/ceph/ceph/pull/17796)
>> > Sage, Josh, Yehuda, Patrick
>> > Is there anything else that needs to go in before we start a second round?
>> >
>> >
>> >
>> > > Regards
>> > > Yan, Zheng
>> > >
>> > > > Best,
>> > > > Abhishek
>> >
>> > --
>> > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>> > the body of a message to majordomo@vger.kernel.org
>> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>

Re: Status of luminous v12.2.1 QE validation

[Patrick Donnelly]

On Mon, Sep 25, 2017 at 10:28 AM, Sage Weil <sweil@redhat.com> wrote:
> On Mon, 25 Sep 2017, Abhishek wrote:
>> On 2017-09-25 11:43, Abhishek wrote:
>> > On 2017-09-25 03:08, Yan, Zheng wrote:
>> > > On 2017/9/24 22:03, Abhishek wrote:
>> > > > On 2017-09-19 23:11, Patrick Donnelly wrote:
>> > > > > On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein <yweinste@redhat.com>
>> > > > > wrote:
>> > > > > > kcephfs - needs analysis - Patrick, Zheng FYI
>> > > > >
>> > > > > http://tracker.ceph.com/issues/21463
>> > > > > http://tracker.ceph.com/issues/21462
>> > > > > http://tracker.ceph.com/issues/21466
>> > > > > http://tracker.ceph.com/issues/21467
>> > > > > http://tracker.ceph.com/issues/21468
>> > > > >
>> > > > > Last 3 look like blockers. Zheng will have more input.
>> > > >
>> > > > Here are the results of a rerun with -k testing,
>> > > > http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/
>> > > > & details http://tracker.ceph.com/issues/21296#note-20
>> > > > There are a few failures which look environmental but a few look like
>> > > > they are related to the changeset (the ones with cache)
>> > > >
>> > > "MDS cache is too large" issue should be fixed by
>> > > https://github.com/ceph/ceph/pull/17922
>> > Ah alright this is a qa suite fix, we can run a the suite with a qa
>> > suite argument just to be sure.
>>
>> alternatively we can just merge this in luminous branch and this gets tested
>> in the final qe run anyway, since we may have to do that since there are some
>> rados and cephfs changes?
>
> Yes, let's do that.
>
> I think all of the other blockers are resolved now?

These whitelist QA fixes would be good to merge to silence spurious
failures for QE:

https://github.com/ceph/ceph/pull/17945
https://github.com/ceph/ceph/pull/17821

Otherwise CephFS looks good to go.

-- 
Patrick Donnelly

Re: Status of luminous v12.2.1 QE validation

[Abhishek]

On 2017-09-25 19:33, Patrick Donnelly wrote:
> On Mon, Sep 25, 2017 at 10:28 AM, Sage Weil <sweil@redhat.com> wrote:
>> On Mon, 25 Sep 2017, Abhishek wrote:
>>> On 2017-09-25 11:43, Abhishek wrote:
>>> > On 2017-09-25 03:08, Yan, Zheng wrote:
>>> > > On 2017/9/24 22:03, Abhishek wrote:
>>> > > > On 2017-09-19 23:11, Patrick Donnelly wrote:
>>> > > > > On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein <yweinste@redhat.com>
>>> > > > > wrote:
>>> > > > > > kcephfs - needs analysis - Patrick, Zheng FYI
>>> > > > >
>>> > > > > http://tracker.ceph.com/issues/21463
>>> > > > > http://tracker.ceph.com/issues/21462
>>> > > > > http://tracker.ceph.com/issues/21466
>>> > > > > http://tracker.ceph.com/issues/21467
>>> > > > > http://tracker.ceph.com/issues/21468
>>> > > > >
>>> > > > > Last 3 look like blockers. Zheng will have more input.
>>> > > >
>>> > > > Here are the results of a rerun with -k testing,
>>> > > > http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/
>>> > > > & details http://tracker.ceph.com/issues/21296#note-20
>>> > > > There are a few failures which look environmental but a few look like
>>> > > > they are related to the changeset (the ones with cache)
>>> > > >
>>> > > "MDS cache is too large" issue should be fixed by
>>> > > https://github.com/ceph/ceph/pull/17922
>>> > Ah alright this is a qa suite fix, we can run a the suite with a qa
>>> > suite argument just to be sure.
>>> 
>>> alternatively we can just merge this in luminous branch and this gets 
>>> tested
>>> in the final qe run anyway, since we may have to do that since there 
>>> are some
>>> rados and cephfs changes?
>> 
>> Yes, let's do that.
>> 
>> I think all of the other blockers are resolved now?
> 
> These whitelist QA fixes would be good to merge to silence spurious
> failures for QE:
> 
> https://github.com/ceph/ceph/pull/17945
> https://github.com/ceph/ceph/pull/17821

About QE validation, my vote is to just go through cephfs & rados suites 
as we have a rados and a few (mostly qe related) cephfs prs that went 
in.

Sage/Yuri/Patrick thoughts?

> 
> Otherwise CephFS looks good to go.

Re: Status of luminous v12.2.1 QE validation

[Sage Weil]

On Tue, 26 Sep 2017, Abhishek wrote:
> On 2017-09-25 19:33, Patrick Donnelly wrote:
> > On Mon, Sep 25, 2017 at 10:28 AM, Sage Weil <sweil@redhat.com> wrote:
> > > On Mon, 25 Sep 2017, Abhishek wrote:
> > > > On 2017-09-25 11:43, Abhishek wrote:
> > > > > On 2017-09-25 03:08, Yan, Zheng wrote:
> > > > > > On 2017/9/24 22:03, Abhishek wrote:
> > > > > > > On 2017-09-19 23:11, Patrick Donnelly wrote:
> > > > > > > > On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein
> > > > <yweinste@redhat.com>
> > > > > > > > wrote:
> > > > > > > > > kcephfs - needs analysis - Patrick, Zheng FYI
> > > > > > > >
> > > > > > > > http://tracker.ceph.com/issues/21463
> > > > > > > > http://tracker.ceph.com/issues/21462
> > > > > > > > http://tracker.ceph.com/issues/21466
> > > > > > > > http://tracker.ceph.com/issues/21467
> > > > > > > > http://tracker.ceph.com/issues/21468
> > > > > > > >
> > > > > > > > Last 3 look like blockers. Zheng will have more input.
> > > > > > >
> > > > > > > Here are the results of a rerun with -k testing,
> > > > > > >
> > > > http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/
> > > > > > > & details http://tracker.ceph.com/issues/21296#note-20
> > > > > > > There are a few failures which look environmental but a few look
> > > > like
> > > > > > > they are related to the changeset (the ones with cache)
> > > > > > >
> > > > > > "MDS cache is too large" issue should be fixed by
> > > > > > https://github.com/ceph/ceph/pull/17922
> > > > > Ah alright this is a qa suite fix, we can run a the suite with a qa
> > > > > suite argument just to be sure.
> > > > 
> > > > alternatively we can just merge this in luminous branch and this gets
> > > > tested
> > > > in the final qe run anyway, since we may have to do that since there are
> > > > some
> > > > rados and cephfs changes?
> > > 
> > > Yes, let's do that.
> > > 
> > > I think all of the other blockers are resolved now?
> > 
> > These whitelist QA fixes would be good to merge to silence spurious
> > failures for QE:
> > 
> > https://github.com/ceph/ceph/pull/17945
> > https://github.com/ceph/ceph/pull/17821
> 
> About QE validation, my vote is to just go through cephfs & rados suites as we
> have a rados and a few (mostly qe related) cephfs prs that went in.
> 
> Sage/Yuri/Patrick thoughts?

Yuri scheduled them last night and they look good (rados was all green :).

I think we're good to go!
sage

Re: Status of luminous v12.2.1 QE validation

[Yuri Weinstein]

Those suite are already done and look good.
fs and kcephfs were reviewed and approved by Patrick.

On Mon, Sep 25, 2017 at 11:20 PM, Abhishek <abhishek@suse.com> wrote:
> On 2017-09-25 19:33, Patrick Donnelly wrote:
>>
>> On Mon, Sep 25, 2017 at 10:28 AM, Sage Weil <sweil@redhat.com> wrote:
>>>
>>> On Mon, 25 Sep 2017, Abhishek wrote:
>>>>
>>>> On 2017-09-25 11:43, Abhishek wrote:
>>>> > On 2017-09-25 03:08, Yan, Zheng wrote:
>>>> > > On 2017/9/24 22:03, Abhishek wrote:
>>>> > > > On 2017-09-19 23:11, Patrick Donnelly wrote:
>>>> > > > > On Tue, Sep 19, 2017 at 1:10 PM, Yuri Weinstein
>>>> > > > > <yweinste@redhat.com>
>>>> > > > > wrote:
>>>> > > > > > kcephfs - needs analysis - Patrick, Zheng FYI
>>>> > > > >
>>>> > > > > http://tracker.ceph.com/issues/21463
>>>> > > > > http://tracker.ceph.com/issues/21462
>>>> > > > > http://tracker.ceph.com/issues/21466
>>>> > > > > http://tracker.ceph.com/issues/21467
>>>> > > > > http://tracker.ceph.com/issues/21468
>>>> > > > >
>>>> > > > > Last 3 look like blockers. Zheng will have more input.
>>>> > > >
>>>> > > > Here are the results of a rerun with -k testing,
>>>> > > >
>>>> > > > http://pulpito.ceph.com/abhi-2017-09-22_19:35:04-kcephfs-luminous-testing-basic-smithi/
>>>> > > > & details http://tracker.ceph.com/issues/21296#note-20
>>>> > > > There are a few failures which look environmental but a few look
>>>> > > > like
>>>> > > > they are related to the changeset (the ones with cache)
>>>> > > >
>>>> > > "MDS cache is too large" issue should be fixed by
>>>> > > https://github.com/ceph/ceph/pull/17922
>>>> > Ah alright this is a qa suite fix, we can run a the suite with a qa
>>>> > suite argument just to be sure.
>>>>
>>>> alternatively we can just merge this in luminous branch and this gets
>>>> tested
>>>> in the final qe run anyway, since we may have to do that since there are
>>>> some
>>>> rados and cephfs changes?
>>>
>>>
>>> Yes, let's do that.
>>>
>>> I think all of the other blockers are resolved now?
>>
>>
>> These whitelist QA fixes would be good to merge to silence spurious
>> failures for QE:
>>
>> https://github.com/ceph/ceph/pull/17945
>> https://github.com/ceph/ceph/pull/17821
>
>
> About QE validation, my vote is to just go through cephfs & rados suites as
> we have a rados and a few (mostly qe related) cephfs prs that went in.
>
> Sage/Yuri/Patrick thoughts?
>
>
>>
>> Otherwise CephFS looks good to go.
>
>