On Wed, Aug 7, 2019 at 2:53 PM Nir Soffer <nsoffer@redhat.com> wrote:
On Wed, Aug 7, 2019 at 1:23 PM Amit Bawer <abawer@redhat.com> wrote:On Wed, Aug 7, 2019 at 11:19 AM Amit Bawer <abawer@redhat.com> wrote:On Tue, Aug 6, 2019 at 5:07 PM Nir Soffer <nsoffer@redhat.com> wrote:On Tue, Aug 6, 2019 at 5:01 PM Amit Bawer <abawer@redhat.com> wrote:On Tue, Aug 6, 2019 at 4:58 PM Nir Soffer <nsoffer@redhat.com> wrote:On Tue, Aug 6, 2019 at 11:27 AM Amit Bawer <abawer@redhat.com> wrote:I have seen some improvement: when I re-trigger the CI per patch I am able to pass or get the actual test errors if any (if not on first try, then on second one).Probably not a very useful information, but I have noticed that when I push 30+ patches at the sameDo not do that, jenkins cannot handle 30 concurrent builds, and is it also bad for reviewers,getting several mails about every patch in your chain, for every rebase.Is there is a way to prevent CI from running per gerrit push (without working on 30 different branches) ?I don't know about such way.A legit option could be adding the Skip CI plugin to jenkins plugins [1]; with that devs can add "[skip ci]" to their commit messages to prevent jenkins from automatically launching CI upon push.Do you want to modify the commit message for every patch to decide if ci should run or not?
I think that having the option to knowingly disable automated CI in some cases is useful. We could always re-trigger when time is right [3].
Another option is to emulate the behaviour in the existing gerrit plugin (guess there is already such one in ovirt's jenkins), for example skipping by a topic regex [2].Not clear how this will help.
If I make a gerrit topic with some name like "my_feature_skip_ci" I can control whether I want to have automated CI for its patches.
When I want to go back to normal I can rename it to "my_feature" and have CI per push as usual.
I think a possible solution can be running only the top patch in a changeset, same way we have in travis,and the same way systems that grab patches from mailing lists work. Every post to gerrit will trigger onebuild, instead of one build per patch in the chain.
That could do as well.
Of course this will allow merging broken patches that are fixed by a later patch in the chain, whichis also not ideal, but it is better given our restricted resources.
We can re-trigger CI manually in this case as part of the verification process.
+Anton Marchukov I have been told you might be familiar with a similar solution.I'm using keeping several small active branches. While you wait for reviews on one topicyou can work on the other branches.time the AWS connection issue arises constantly._______________________________________________On Sun, Aug 4, 2019 at 4:49 PM Eyal Edri <eedri@redhat.com> wrote:This was reported already and AFAIK its a network issue between AWS and PHX which is still being investigated.Evgheni, any insights or update on the issue? should we involve debugging from amazon side?On Sun, Aug 4, 2019 at 4:46 PM Amit Bawer <abawer@redhat.com> wrote:_______________________________________________Hi,CI seems to fail constantly for unavailable remote gerrit repository.Example can be seen here:
Devel mailing list -- devel@ovirt.org
To unsubscribe send an email to devel-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/devel@ovirt.org/message/AHPHUZAABAQNWEMD2JQ6WARHJRDTYCPI/
--phone: +972-9-7692018
irc: eedri (on #tlv #rhev-dev #rhev-integ #cp-devel)
Devel mailing list -- devel@ovirt.org
To unsubscribe send an email to devel-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/devel@ovirt.org/message/W6DUMIUSN5DPUVGUFUNHF2ZALB5I4JPZ/
