Summary: | Hudson builds failing with "insufficient memory for the Java Runtime Environment to continue" | ||
---|---|---|---|
Product: | Community | Reporter: | Sam Davis <sam.davis> |
Component: | CI-Jenkins | Assignee: | CI Admin Inbox <ci.admin-inbox> |
Status: | RESOLVED FIXED | QA Contact: | |
Severity: | normal | ||
Priority: | P1 | CC: | denis.roy, webmaster |
Version: | unspecified | ||
Target Milestone: | --- | ||
Hardware: | PC | ||
OS: | Windows 7 | ||
Whiteboard: |
Description
Sam Davis
2016-05-25 19:21:55 EDT
My guess is that the host for you HIPP is really busy(I'm seeing some OOM events in the logs). About the only thing I can suggest is moving your HIPP to another host(we have a couple of spare slots). If you'd like me to do that I'll need to shut your HIPP down for a few hours(depending on disk space usage) while I transfer the instance, so we'll need to pick a time. -M. Yes, please move it to another host as soon as you can. We can deal with HIPP being down for a few hours. Ok I've finished the move. -M. Thanks. But I think there's a problem with XVNC. The console output has a bunch of errors like "** (java:17420): WARNING **: Could not open X display" and the Eclipse log says "org.eclipse.swt.SWTError: No more handles [gtk_init_check() failed]." https://hudson.eclipse.org/mylyn/job/mylyn-commons-nightly/43/consoleFull In another build, we have: Warning: VNC extension does not support -reset, terminating instead. Use -noreset to prevent termination. Unable to init server: Could not connect: Connection refused An error has occurred. See the log file /jobs/genie.mylyn/gerrit-mylyn-tasks/workspace/org.eclipse.mylyn.bugzilla.tests/target/work/data/.metadata/.log. The log file contains the same message about no handles. Missing new Xvnc options. Fixed. -M. Thanks, that seems to have worked. But now it seems that the builds cannot access /usr/bin/sign: [exec] + /usr/bin/sign /home/data/httpd/download-staging.priv/tools/mylyn/signing/mylyn/3.20.0-SNAPSHOT/site.zip nomail /home/data/httpd/download-staging.priv/tools/mylyn/signing/mylyn/3.20.0-SNAPSHOT/output [exec] /jobs/genie.mylyn/mylyn-3.20.x-release/workspace/org.eclipse.mylyn/org.eclipse.mylyn-site/pack-and-sign/sign-and-wait.sh: line 42: /usr/bin/sign: No such file or directory https://hudson.eclipse.org/mylyn/job/mylyn-3.20.x-release/10/console Also, it seems this host may have a buggy version of wget? [1] We keep getting this failure in https://hudson.eclipse.org/mylyn/job/mylyn-tasks-nightly/ + wget -O org.eclipse.mylyn.commons.stamp https://hudson.eclipse.org/mylyn/job/mylyn-commons-nightly/lastSuccessfulBuild/artifact/org.eclipse.mylyn.commons.stamp --2016-05-27 17:05:12-- https://hudson.eclipse.org/mylyn/job/mylyn-commons-nightly/lastSuccessfulBuild/artifact/org.eclipse.mylyn.commons.stamp Resolving proxy.eclipse.org (proxy.eclipse.org)... 172.30.206.220 Connecting to proxy.eclipse.org (proxy.eclipse.org)|172.30.206.220|:9898... connected. Proxy tunneling failed: Bad RequestUnable to establish SSL connection. [1] https://www.reddit.com/r/sysadmin/comments/34a9sz/what_the_hell_wget/ Any update on this? We can work around the wget issue by disabling fingerprinting but we can't do a release build if we can't access /usr/bin/sign. Raising to P1 because the deadline for RC3 +3 builds is tomorrow. We're on this, Sam. Sorry for the delay. I think many projects have moved to the cbi signing maven plugin. https://www.eclipse.org/cbi/sitedocs/index.html Fixed, sorry about that. -M. Thanks very much. The release build is working now, and I can work around wget being broken (comment 8) by using curl instead. And it seems as though moving hosts fixed the insufficient memory problem. |