[
https://issues.apache.org/jira/browse/HBASE-20364?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16437939#comment-16437939
]
Sean Busbey commented on HBASE-20364:
-------------------------------------
{quote}
to make sure i understand... we write a default failure and then when we finish
we write the real result. but if we don't finish and time out then we still
have the assumed failure message written?
{quote}
If we don't finish in a way that still allows post actions, but yes that's
essentially correct.
{quote}
do we know that this will only happen in case of timeout? will sometime like
OOM also cause the stage to fail without writing a message or will that get a
known failure state? trying to understand if we can write a more specific
initial fail message
{quote}
No it's not just timeout. we could, for example, transiently run into a
resource limit while trying to use the "gather machine stats" script. That
would fail the stage but still run post, which would have us use the default
message.
We could get into updating the message incrementally as we go, but I'm
skeptical of incremental benefit beyond "we hit an unexpected failure of the
test infra."
> nightly job gives old results or no results for stages that timeout on SCM
> --------------------------------------------------------------------------
>
> Key: HBASE-20364
> URL: https://issues.apache.org/jira/browse/HBASE-20364
> Project: HBase
> Issue Type: Bug
> Components: test
> Affects Versions: 3.0.0
> Reporter: Sean Busbey
> Assignee: Sean Busbey
> Priority: Critical
> Attachments: HBASE-20364.0.patch
>
>
> seen in the branch-2.0 nightly report for HBASE-18828:
>
> {quote}
> Results for branch branch-2.0
> [build #143 on
> builds.a.o|https://builds.apache.org/job/HBase%20Nightly/job/branch-2.0/143/]:
> (x) *\{color:red}-1 overall\{color}*
> ----
> details (if available):
> (/) \{color:green}+1 general checks\{color}
> -- For more information [see general
> report|https://builds.apache.org/job/HBase%20Nightly/job/branch-2.0/140//General_Nightly_Build_Report/]
>
> (/) \{color:green}+1 jdk8 hadoop2 checks\{color}
> -- For more information [see jdk8 (hadoop2)
> report|https://builds.apache.org/job/HBase%20Nightly/job/branch-2.0/143//JDK8_Nightly_Build_Report_(Hadoop2)/]
> (/) \{color:green}+1 jdk8 hadoop3 checks\{color}
> -- For more information [see jdk8 (hadoop3)
> report|https://builds.apache.org/job/HBase%20Nightly/job/branch-2.0/143//JDK8_Nightly_Build_Report_(Hadoop3)/]
>
> {quote}
>
> -1 for the overall build was correct. build #143 failed both the general
> check and the source tarball check.
>
> but in the posted comment, we get a false "passing" that links to the general
> result from build #140. and we get no result for the source tarball at all.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
