I agree there're ideas how to handle/improve it that, but I'd like to point out that trying to do so, we may worsen that instead. I don't think ratio of current rate of random failures vs possible benefits vs possible risk calls for any actions, so let's have this ticket to collect stats of such failures, so far it was very low.
** Changed in: linaro-android-mirror Importance: Undecided => Low ** Project changed: linaro-android-mirror => linaro-android-build-tools -- You received this bug notification because you are a member of linaro- infrastructure-drivers, which is the registrant for linaro-android- mirror. https://bugs.launchpad.net/bugs/855050 Title: Needs to cope with transient failure better Status in Linaro Android Build Tools: New Bug description: Hi Filing this here so it's on a Linaro project, but we may fix it elsewhere. https://android-build.linaro.org/jenkins/job/linaro-android_staging- omap4460/2/parsed_console/? failed with error: Ref refs/remotes/origin/linaro_android_2.3.5 is at c1c999294e24a97466b4a863a96dc09d8c389088 but expected 67102998c3a453a1bd173ed3eec4bfafe077ba7a From git://android.git.linaro.org/platform/manifest ! 6710299..c1c9992 linaro_android_2.3.5 -> origin/linaro_android_2.3.5 (unable to update local ref) From git://android.git.linaro.org/platform/manifest 6710299..c1c9992 linaro_android_2.3.5 -> linaro_android_2.3.5 From git://android.git.linaro.org/platform/external/busybox + 17afabe...5c6ba6c master -> master (forced update) Initializing project kernel/omap-omapzoom ... fatal: The remote end hung up unexpectedly fatal: protocol error: bad pack header error: Cannot fetch kernel/omap-omapzoom error: Exited sync due to fetch errors which was apparently the connection with android.git.linaro.org closing prematurely. The next build (immediately following) worked fine. It may be that transient errors are handled, and this downtime just exceeded a threshold, but it's likely that they aren't handled at all. In order to have a reliable service short-term transient errors should be transparently handled. We could do this by running repo again if it fails. Given that it takes 10 minutes to do nothing, that may be rather expensive if it's a 10 second outage. We could have repo re-run git if it fails, which would be more direct. Or git could try reconnecting. Or all of the above. Thanks, James To manage notifications about this bug go to: https://bugs.launchpad.net/linaro-android-build-tools/+bug/855050/+subscriptions _______________________________________________ Mailing list: https://launchpad.net/~linaro-release Post to : [email protected] Unsubscribe : https://launchpad.net/~linaro-release More help : https://help.launchpad.net/ListHelp
