7.x Stops Checking For Pending Changes

I have several configurations all pointing to the same SVN server.  We just committed a change an hour ago and we've yet to see ny pending changes to appear or any of our build configurations run.  

This is a TC setup that has been running for over a year now and only in the last few months (since upgrading from TC 6.x to 7.x) have we started to get periods where pending changes are not detected or executed.  In one case a single build configuration failed to run for 1 whole week...  it didn't work again until i rebooted the server.

Checking the teamcity-vcs.log there are no errors of any kind.  Looking at the last line looks like this...

[2012-08-15 13:46:19,581]   INFO [1, parent id=1}] -      jetbrains.buildServer.VCS - Finish collecting changes for svn: https://sickhead.jira.com/svn/BC/trunk/Code {instance id=11, parent id=1} from state SingleVersionRState{5877_2012/08/15 13:45:59 -0500 at Wed Aug 15 13:46:19 CDT 2012} to state SingleVersionRState{5877_2012/08/15 13:46:19 -0500 at Wed Aug 15 13:46:19 CDT 2012}; 0 changes collected 0 changes reported, time spent: 562ms

Note this is over 1 hour ago now... so for 1 whole hour the server has stopped checking for changes.

I cannot spend my days rebooting the build server to make it "wake up" and decide to check for changes.  How do i fix this???

- Tom

4 comments
Comment actions Permalink

Finally... ***2 hours**** later TC starts getting pending changes.

In the log i see this...

[2012-08-15 15:46:39,235]   WARN [1, parent id=1}] -      jetbrains.buildServer.VCS - Exception occured on attempt to collect changes by combined checkout rule: =>
 (jetbrains.buildServer.vcs.VcsException: org.tmatesoft.svn.core.SVNException: svn: E175002: timed out waiting for server
svn: E175002: OPTIONS request failed on '/svn/BC/trunk/Code'), will try to collect changes for each rule separately 
[2012-08-15 15:46:40,967]   INFO [1, parent id=1}] -      jetbrains.buildServer.VCS - Finish collecting changes for svn: https://sickhead.jira.com/svn/BC/trunk/Code {instance id=11, parent id=1} from state SingleVersionRState{5877_2012/08/15 13:46:19 -0500 at Wed Aug 15 13:46:39 CDT 2012} to state SingleVersionRState{5877_2012/08/15 13:46:39 -0500 at Wed Aug 15 13:46:39 CDT 2012}; 0 changes collected 0 changes reported, time spent: 2h:0m:01s,964ms 
[2012-08-15 15:46:43,776]   INFO [rting build bt9] -      jetbrains.buildServer.VCS - Finish collecting changes for svn: https://sickhead.jira.com/svn/BC/trunk/Code {instance id=11, parent id=1} from state SingleVersionRState{5877_2012/08/15 13:46:39 -0500 at Wed Aug 15 15:46:41 CDT 2012} to state SingleVersionRState{5878_2012/08/15 15:46:41 -0500 at Wed Aug 15 15:46:41 CDT 2012}; 1 changes collected 1 changes reported, time spent: 2s,013ms 


So according to the log it too 2 full hours to "timeout" waiting for the server to reply for change requests.  Mind you the SVN server was available the whole time...  it wasn't down as we have been commiting changes and checking SVN logs during this period.

Where is the timeout value i can tweak to get it to stop waiting on the server in a reasonable amount of time?  Say 2 minutes... not 2 hours.

- Tom
0
Comment actions Permalink
Anybody home?
0
Comment actions Permalink
Amost a week...  no Jetbrains support as of yet.
0
Comment actions Permalink

Tom,

I apologize for the delay.

What exact version of TeamCity do you use? Some related fixes were delivered in 7.1 (e.g., TW-20649 and TW-22125)

Also, please read the comments here: TW-21789. The issue was accidentally resolved by VisualSVN configuration.

0

Please sign in to leave a comment.