Why is 'Checking for Changes' taking far longer after 6.5.5 upgrade?

Noticing that after I upgraded to 6.5.5 my builds are sitting on 'Checking for Changes' for several minutes. I'm using kiln version 2.6.41 (mercurial).  Cloning the repos directly on the build server seems to be about the normal speed.  ~4-5 mins checking for changes in teamcity, 30 secs to clone when done manually via tortisehg.

Tried deleting everything under .buildserver/caches/mercurial but that didnt help.

I'm using server side checkout.  Tried switching to agent checkout, but that didnt help either.

Any thoughts? Can provide configuration details as needed...

Thanks!

3 comments
Comment actions Permalink

Yes, post teamcity-vcs.log please.

Michael

0
Comment actions Permalink

tons of these warnings due to hg.exe moving from "Program Files (x86)" to just "Program Files"

[2011-11-03 11:46:05,643]   WARN [utor 4 {id=182}] -      jetbrains.buildServer.VCS - Error while loading changes for root mercurial: http://XXXXX.com/kiln/Repo/XXXXXXXXX {instance id=182, parent id=2}, cause: '"C:\Program Files (x86)\TortoiseHg\hg.exe" clone --pull -U --uncompressed http://USER:******@XXXXX.com/kiln/Repo/XXXXXXXXX hg_6944142928962812896' command failed.
exception: Cannot run program ""C:\Program Files (x86)\TortoiseHg\hg.exe"" (in directory "C:\.BuildServer\system\caches\mercurial"): CreateProcess error=2, The system cannot find the file specified



i fixed most of the configs, missed a few

This does not seem like it would effect the build configurations that were already fixed to use the correct hg.exe path though....

More details from build log of same project:

Before upgrade:
[11:06:08]: Checking for changes
[11:06:11]: Clearing temporary directory: E:\BuildAgent\temp\buildTmp

After upgrade:
[14:38:02]: Checking for changes

[14:44:32]: Clearing temporary directory: E:\BuildAgent\temp\buildTmp



Info messages for this:
[2011-11-03 12:44:31,519]   INFO [rmal executor 5] -      jetbrains.buildServer.VCS - Finished collecting changes for build configurations [Metrics :: Web App CI {id=bt4}], time spent: 6m:29s,078ms
[2011-11-03 12:44:41,587]   INFO [rmal executor 1] -      jetbrains.buildServer.VCS - Requesting patch: root=mercurial: http://XXXXXXX/kiln/Repo/YYYYYYYY {instance id=278, parent id=2}, cleanPatch=true, prevVersion=null, newVersion=118:09226b9ef59e, buildType=Metrics :: Web App CI {id=bt4}, buildId=8353
[2011-11-03 12:45:15,400]   INFO [rmal executor 1] -      jetbrains.buildServer.VCS - Done requesting patch for root mercurial: http://XXXXXXX/kiln/Repo/YYYYYYYY {instance id=278, parent id=2} cleanPatch = true, prevVersion=null, newVersion=118:09226b9ef59e, buildId = 8353; took 33 sec 809 msec
[2011-11-03 12:45:37,197]   INFO [rmal executor 8] -      jetbrains.buildServer.VCS - Requesting patch: root=mercurial: http://XXXXXXX/kiln/Repo/YYYYYYYY {instance id=278, parent id=2}, cleanPatch=true, prevVersion=null, newVersion=118:09226b9ef59e, buildType=Metrics :: Web App CI {id=bt4}, buildId=8353
0
Comment actions Permalink

so it seems like TeamCity was not at fault here

i talked with the kiln support team and they indicated that having teamcity use tortisehg version 2.1.2 was likely causing the perf issue

i downgraded all the teamcity servers back to the previous hg version with kiln still at the new version and the performance went back to normal

ex

[11:55:19]: Checking for changes
[11:55:26]: Clearing temporary directory: E:\BuildAgent\temp\buildTmp
0

Please sign in to leave a comment.