Bugzilla – Full Text Bug Listing
| Summary: | Intermittent build failures -eclipse-master-${buildId}.zip doesn't exist | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Eclipse Project] Platform | Reporter: | Kim Moir <kim.moir> | ||||||||
| Component: | Releng | Assignee: | Kim Moir <kim.moir> | ||||||||
| Status: | RESOLVED FIXED | QA Contact: | |||||||||
| Severity: | critical | ||||||||||
| Priority: | P3 | CC: | aniefer, daniel_megert, sonia_dimitrov | ||||||||
| Version: | 3.2 | ||||||||||
| Target Milestone: | --- | ||||||||||
| Hardware: | PC | ||||||||||
| OS: | Windows XP | ||||||||||
| Whiteboard: | |||||||||||
| Bug Depends on: | 172935 | ||||||||||
| Bug Blocks: | |||||||||||
| Attachments: |
|
||||||||||
|
Description
Kim Moir
Last week, I increased maxmemory="512m" in the pack200 stage in eclipse/helper.xml. The problem is very intermittent so it is difficult to reproduce. I'm closing this bug, since I added more memory to the pack200 arguments, we have not seen this intermittent failure. Happened again with N20060727-0010. I think the memory setting needs to be set in the extras/pack200 script, ie. @pack200@ -E4 -J-Xms512m $* Closing, hasn't occured since memory changes. Happened again this morning. Andrew advised on implementing additionally loggin in the pack200 section
<!--pack200-->
<java jar="${eclipse.home}/startup.jar"
fork="true"
timeout="10800000"
jvm="${java15-home}/bin/java"
failonerror="true"
maxmemory="768m"
error="${buildDirectory}/errorlog.txt"
dir="${buildDirectory}"
output="${buildDirectory}/jarprocessorlog.txt">
<jvmarg value="-Dorg.eclipse.update.jarprocessor.pack200=${eclipse.build.configs}/../../extras"/>
<arg line="-consolelog -application org.eclipse.update.core.siteOptimizer"/>
<arg line="-jarProcessor -verbose -outputDir ${buildLabel} -processAll -pack ${repack} ${packtmp}/${archiveName}"/>
</java>
From what we can see, the process exits when packing a single jar. The jar that causes the process to exit differs from build to build.
Created attachment 58271 [details]
logs from failed packing process
It happened on a a test build this morning. Andrew, I'm attaching the logs.
Since our signing server was down, both packing and repacking occurred on our build machine.
The source of the errorlog.txt and jarprocessorlog.txt can bee seen from the ant snippet above.
The packlog.txt is generated by calling from the output of the pack200 script that calls the pack200 executable from the vm.
#!/bin/sh
/builds/N200702051147/jdk/linuxppc/ibm-java2-ppc-50/jre/bin/pack200 -v -l packlog.txt -E4 -J-Xmx768m $*
pack200retCode=$?
if [ $pack200retCode != 0 ]
then
echo "Build failed (error code $retCode)."
exit -1
fi
Created attachment 58283 [details]
patch to update.core
Patch to update.core for the JarProcessor to handle this file not found more gracefully.
Created attachment 58284 [details]
update.core jar with changes
New update.core jar containing the attached changes to jarprocessor.
Thanks Andrew, I will run a test build tomorrow. I didn't have time today with all our our build woes. test build successful, jar released into basebuilder for midnight build. Looks like this is fixed. I haven't had it happen since I released the fix. Sonia, when I update basebuilder with M5, I'll update it with this update.core jar as Andrew's patch was not released to that plugin for M5. closing |