-
Bug
-
Resolution: Unresolved
-
Major
See second comment for updated issue cause.
I have configured project, where in csm I have Additional Behaviours -> Pooling ignores commits in certain paths with Included Regions pointing to one specific file. I have there also additional behaviour "Clean before checkout".
When post-receive hook triggers change, everything works as expected.
Pool SCM schedule is set to "H/1 * * * *" (each hour).
Yesterday I have found that build has been started in night despite no changes was made to given file. In pool log for given build is this:
NOTE: In second comment https://issues.jenkins-ci.org/browse/JENKINS-38722?focusedCommentId=272666&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-272666 is actual pooling log of problematic build triggering. This build has been started by something else (I do not know by what), but it is very similar to two referenced similar issues JENKINS-20389 and JENKINS-23606.
Started on Oct 4, 2016 3:11:54 PM Polling SCM changes on master Using strategy: Default [poll] Last Built Revision: Revision 5eb3f4c485189f8151d0fe777a31bcfc48adb09e (origin/master) > git rev-parse --is-inside-work-tree # timeout=10 Fetching changes from the remote Git repositories > git config remote.origin.url ssh://git@dev.asdf.sk/~/salt.git/ # timeout=10 Cleaning workspace > git rev-parse --verify HEAD # timeout=10 Resetting working tree > git reset --hard # timeout=10 > git clean -fdx # timeout=10 Fetching upstream changes from ssh://git@dev.asdf.sk/~/salt.git/ > git --version # timeout=10 > git fetch --tags --progress ssh://git@dev.asdf.sk/~/salt.git/ +refs/heads/*:refs/remotes/origin/* Polling for changes in > git rev-parse origin/master^{commit} # timeout=10 > git log --full-history --no-abbrev --format=raw -M -m --raw 5eb3f4c485189f8151d0fe777a31bcfc48adb09e..abf529a33a519c6c712a419b19ac969d20896d6f # timeout=10 Done. Took 0.46 sec Changes found
In console output is this:
Started by an SCM change [EnvInject] - Loading node environment variables. Building on master in workspace /var/lib/jenkins/workspace/autodeploy-stagebackend > git rev-parse --is-inside-work-tree # timeout=10 Fetching changes from the remote Git repository > git config remote.origin.url ssh://git@dev.asdf.sk/~/salt.git/ # timeout=10 Cleaning workspace > git rev-parse --verify HEAD # timeout=10 Resetting working tree > git reset --hard # timeout=10 > git clean -fdx # timeout=10 Fetching upstream changes from ssh://git@dev.asdf.sk/~/salt.git/ > git --version # timeout=10 > git fetch --tags --progress ssh://git@dev.asdf.sk/~/salt.git/ +refs/heads/*:refs/remotes/origin/* > git rev-parse origin/master^{commit} # timeout=10 Checking out Revision abf529a33a519c6c712a419b19ac969d20896d6f (origin/master) > git config core.sparsecheckout # timeout=10 > git checkout -f abf529a33a519c6c712a419b19ac969d20896d6f > git rev-list 5eb3f4c485189f8151d0fe777a31bcfc48adb09e # timeout=10 No emails were triggered. [autodeploy-stagebackend] $ /bin/bash -xe /tmp/hudson3936035578572855045.sh ....
I think that pool log may sometime indicates ("No workspace is available, so can’t check for updates.") use of git ls-remote (https://wiki.jenkins-ci.org/display/JENKINS/Git+Plugin#GitPlugin-FastRemotePolling), So from morning I have added additional behaviour "force pooling using workspace". I hope it will help a bit tracking issue down.
What has been done in git:
5eb3f4c485189f8151d0fe777a31bcfc48adb09e last successfully build commit. It included file mentioned in included regions. It is commit with message "PRH-1542 New version of some app for stage"
abf529a33a519c6c712a419b19ac969d2089 merge commit which merges given commit (rebase was not used when git pull was done, thus merge was done)
Output of git log --full-history --no-abbrev --format=raw -M -m --raw 5eb3f4c485189f8151d0fe777a31bcfc48adb09e..abf529a33a519c6c712a419b19ac969d20896d6f
commit abf529a33a519c6c712a419b19ac969d20896d6f (from 0a495d7f6b5a8db46edd2ef72a1bf9753ed87a84) tree 7f00c97840e8d6243362976310301c1c9dc70c7d parent 0a495d7f6b5a8db46edd2ef72a1bf9753ed87a84 parent 5eb3f4c485189f8151d0fe777a31bcfc48adb09e author L X <lx@asdf.sk> 1475586630 +0200 committer L X <lx@asdf.sk> 1475586630 +0200 Merge branch 'master' of ssh://dev.asdf.sk/~/salt * 'master' of ssh://dev.asdf.sk/~/salt: PRH-1542 New version of bonus-data for stage. :100644 100644 3861e9a150a7f4d4713090575aeb04b85ec304cc c3fba55d424f5ea53103a088c30860438adf27ec M pillar/backend/stage/init.sls commit abf529a33a519c6c712a419b19ac969d20896d6f (from 5eb3f4c485189f8151d0fe777a31bcfc48adb09e) tree 7f00c97840e8d6243362976310301c1c9dc70c7d parent 0a495d7f6b5a8db46edd2ef72a1bf9753ed87a84 parent 5eb3f4c485189f8151d0fe777a31bcfc48adb09e author L X <lx@asdf.sk> 1475586630 +0200 committer L X <lx@asdf.sk> 1475586630 +0200 Merge branch 'master' of ssh://dev.asdf.sk/~/salt * 'master' of ssh://dev.asdf.sk/~/salt: PRH-1542 New version of some app for stage. :100644 100644 9ec0e5491222d6011af2859ccd2d476d02b84e2d bd9f2f6b08834e5196b6913b7cc1b07db8f94a4b M pillar/centrala/production/asdf.sls commit 0a495d7f6b5a8db46edd2ef72a1bf9753ed87a84 tree 9ce7c6cb816c5762e859a7932049d8b00feb5b9f parent 936b93fcb21501997450561a3c08b404dc610344 author L X <lx@asdf.sk> 1475586451 +0200 committer L X <lx@asdf.sk> 1475586451 +0200 db_bonusy connect :100644 100644 9ec0e5491222d6011af2859ccd2d476d02b84e2d bd9f2f6b08834e5196b6913b7cc1b07db8f94a4b M pillar/centrala/production/asdf.sls
So I think that merge (wchich have not changed file since last build) trigered build. My suggestion would be to use git diff instead of git log for changes detection:
git diff 5eb3f4c485189f8151d0fe777a31bcfc48adb09e..abf529a33a519c6c712a419b19ac969d20896d6f
diff --git a/pillar/centrala/production/asdf.sls b/pillar/centrala/production/asdf.sls index 9ec0e54..bd9f2f6 100644 --- a/pillar/centrala/production/asdf.sls +++ b/pillar/centrala/production/asdf.sls @@ -1,2 +1,9 @@ # vim: expandtab ts=2 ft=yaml autoindent indentexpr= fdm=marker fmr=/*,*/ +redukcia: + db: + db_bonusy: + dsn: DBI:mysql:bonusy;mysql_socket=/var/lib/mysql/asdf.sock;mysql_read_default_group=db_latin2;mysql_read_default_file=/usr/roft/etc/my.cnf.txt + user: jankoMrkva + passwd: nourčite +
PS: I still do not know, why build was triggered at night/morning, because given merge was done at lunch time and so from given time, there should be done plenty of scheduled git pull actions. Also on build time, there were no git commit. It is possible
- is related to
-
JENKINS-20389 Merge commits trigger uneccessary builds when Include/Exclude regions are defined
- Open
-
JENKINS-23606 Git Polling Jobs with Included Regions being triggered unnecessarily with merge commits
- Open