Apache geode nightly failures

[Jason Huynh <ja...@apache.org>]

Has anyone been getting emails for the nightly builds and failures (My last
build email was 1027 and I think the latest is 1030)?  It looks like the
last few have been a mess with Out of Memory exceptions and I think they
were all ran on H34.  Should that machine be black listed?  If so, would
someone with permissions be able to do so?

Re: Apache geode nightly failures

[Dick Cavender <dc...@pivotal.io>]

I just blacklisted H34 but some 12 days back the nightly ran fine on H34
after blacklisting H12 and H14 as not being reliable. Now H34 has become
the same.

This is the complete list of specific machines and the same probably
happened when we added H2.
&&!H2&&!H12&&!H14&&!H34

Smells like there are project/workspaces out there tainting these
machine(s) when they run or complete. Let's see what machine the next
nightly build lands while pushing towards getting some reliable history on
concourse and transition off Jenkins.

Reminder that the geode concourse CI can be viewed here:
https://concourse.apachegeode-ci.info/teams/main/pipelines/develop

-Dick


On Fri, Dec 1, 2017 at 9:29 AM, Jason Huynh <ja...@apache.org> wrote:

> Has anyone been getting emails for the nightly builds and failures (My last
> build email was 1027 and I think the latest is 1030)?  It looks like the
> last few have been a mess with Out of Memory exceptions and I think they
> were all ran on H34.  Should that machine be black listed?  If so, would
> someone with permissions be able to do so?
>