Closed Bug 473066 Opened 16 years ago Closed 16 years ago

20090111 Windows Minefield trunk nightly build failed generating the complete .mar file

Categories

(Release Engineering :: General, defect, P2)

x86
Windows XP

Tracking

(Not tracked)

RESOLVED WORKSFORME

People

(Reporter: wgianopoulos, Assigned: nthomas)

Details

Today's (20090111) nightly win32 build failed during the creation of the full .mar file with the error message: mv: cannot stat`../../dist/firefox.work/output.mar': No such file or directory
Summary: 20090111 Windows nightly build failed generating the complete .mar file → 20090111 Windows Minefield trunk nightly build failed generating the complete .mar file
This was on moz2-win32-slave11, which luckily didn't clobber the nightly dir on subsequent builds. Slave is disconnected for now.
Assignee: nobody → nthomas
Status: NEW → ASSIGNED
Priority: -- → P2
This is kinda weird and I'm inclined to think it's a one-off glitch (having not seen it before). * The log for the buildbot step ends ignoring remove instruction for directory: searchplugins/ ignoring remove instruction for directory: extensions/talkback@mozilla.org/ mv: cannot stat `../../dist/firefox.work/output.mar': No such file or directory make: Leaving directory `/e/builds/moz2_slave/mozilla-central-win32-nightly/build/obj-firefox/tools/update-packaging' program finished with exit code 0 * if you look at a normal log then it's exactly the same except for the cannot stat error * the Makefile in tools/update-packaging is calling make_full_update.sh, which ends # Append remove instructions for any dead files. append_remove_instructions "$targetdir" >> $manifest $BZIP2 -z9 "$manifest" && mv -f "$manifest.bz2" "$manifest" eval "$MAR -C \"$workdir\" -c output.mar $targetfiles" mv -f "$workdir/output.mar" "$archive" # cleanup rm -fr "$workdir" * the "ignoring remove instruction" output comes from append_remove_instructions * output.mar is created by the call to $MAR * the "cannot stat" comes from the mv call * the rm should not be starting before the mv has finished, (WHOOP WHOOP WHOOP, LOSS OF CAUSALITY ALERT). RESOLVED WTF. Actually, I just triggered another m-c nightly on the original code revision, and moved on with my day.
Status: ASSIGNED → RESOLVED
Closed: 16 years ago
Resolution: --- → WORKSFORME
Do I want to ask what a one-off glitch means? Non-determinism in the build infrastructure makes me nervous! :)
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.