Last week they suggested to use a sudo-enabled environment (by modifying our travis.yml) as they have more memory available. This should only buy us a few minutes though.

I haven't heard from them since, and they never provided an explanation for increased
build times. According to them nothing changed on their side.

It is not a problem on our side though, that's a fact by now. Branches that haven't been
modified, like the release-1.2 branch, now also time out.

That said, if both travis and we haven't changed anything, maybe Apache/Infra had
anything to do with it?

On 22.06.2017 12:12, Robert Metzger wrote:
Hi Chesnay,
Thanks a lot for taking care of the issue!
Did you get any insights by talking to the travis support about this?

On Tue, Jun 13, 2017 at 5:07 PM, Chesnay Schepler <ches...@apache.org>
wrote:

Hello,

there currently isn't a good workaround for contributors.

I'm currently reaching out to Travis support as to whether something about
their infrastructure was changed recently.
It seems unlikely that we didn't notice a commit that adds 50 minutes to
the build times.

Regards,
Chesnay


On 13.06.2017 16:56, Jark Wu wrote:

Hi devs,

I used to check whether a PR can be merged by using Travis CI on my own
repo branch. But recently I find it doesn't work. Such as this branch [1]
,
   Travis CI failed when reached 50 minutes limit. I find that flink repo
can
run more than 1 hr 30 min. Do you know how to increase the time limit of
Travis for my repo?   Or how do you check a PR can be merged currently?

[1] https://travis-ci.org/wuchong/flink/builds/242259542


Thanks,
Jark Wu



Reply via email to