Re: [Gluster-infra] Can I update /opt/qa for NetBSD?

[Nigel Babu]

Hello folks,

I've been thinking about a different approach here. I'd like to fork the
build.sh/regression.sh/smoke.sh scripts for NetBSD until we unify them. This
seems to be the approach with most gains at this point.

This way, I can make smaller bite sized changes that can be tested in
production (as a separate test job). This also means what's in the repo is
what's currently running.

So, I'd have netbsd-smoke.sh, netbsd-regression.sh, and netbsd-build.sh for a
while. I will work on merging the changes into one script. Watch out for me
nagging you for reviews :)

Speaking of harassing for reviews: 
https://github.com/gluster/glusterfs-patch-acceptance-tests/pull/50

Practically, I only need someone to look at one line: 
https://github.com/gluster/glusterfs-patch-acceptance-tests/pull/50/files#diff-db1659608ae45b2644416de2b9a76d3cR8

--
nigelb
___
Gluster-infra mailing list
Gluster-infra@gluster.org
http://www.gluster.org/mailman/listinfo/gluster-infra

Re: [Gluster-infra] Can I update /opt/qa for NetBSD?

[Emmanuel Dreyfus]

Nigel Babu  wrote:

> I'll definitely appreciate any feedback you can have in terms
> of code when it's ready for review.

No problem. But regresson infrastructure will catch any issue better
that I would, anyway. :-)

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
m...@netbsd.org
___
Gluster-infra mailing list
Gluster-infra@gluster.org
http://www.gluster.org/mailman/listinfo/gluster-infra

Re: [Gluster-infra] Can I update /opt/qa for NetBSD?

[Emmanuel Dreyfus]

Nigel Babu  wrote:

> * A bunch of cleanup scripts are in build.sh. I think they can move into the
>   Jenkins job itself. That's what we do for linux boxes.

Feel free to do it, but is there a benefit? If it is not broken, do not
fix it...

> * Some test files have been made to 0644. Is this relevant or are these
>   accidental changes?

Probable accidental change.

> * Can the Python path and stuff be declared before the build.sh script is
>   called in the Jenkins script? I don't know if this will work, I'll have to
>   test.

You need to run build.sh so that configure is invoked and env.rc is
created with appropriate @BUILD_PYTHON_SITE_PACKAGES@ set
(see tests/env.rc.in)

> * I think we now have a standard way of skipping tests in the test runner
>   itself rather than deleting the tests from the checkout. If not, I'll drive
>   these fixes.

It may not scale when you want to skip all the bugs directory.

> * The check for whether there's two jobs assigned to the same machine can
> be controlled from Jenkins and I plan to do that, so we can probably
> remove that code as well.

That seems better, as the current setting bugs when a job is manually
cancelled and retriggered (here is a point to fix!)

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
m...@netbsd.org
___
Gluster-infra mailing list
Gluster-infra@gluster.org
http://www.gluster.org/mailman/listinfo/gluster-infra

[Gluster-infra] Can I update /opt/qa for NetBSD?

[Nigel Babu]

Hello folks,

The /opt/qa folder on NetBSD nodes are on quite an old commit[1]. This means
the URL to the logs and cores do not work. The script in master generates the
correct URL. Is it safe to update this repo to the latest git master?

[1]: 
https://github.com/gluster/glusterfs-patch-acceptance-tests/commit/931157d3db4c48a985c3874c2d4ccf2ebb2fbd5f

--
nigelb
___
Gluster-infra mailing list
Gluster-infra@gluster.org
http://www.gluster.org/mailman/listinfo/gluster-infra