Hi
with the Queues it is guranteed to work, with the initial seconds, it will work on startup, but after a while there will be a time that 2 projects will fire again at the same time. you can combine the initial seconds with the queues if you want. with kind regards Ruben Willems On Mon, Jul 27, 2009 at 10:18 AM, Michael <[email protected]> wrote: > > Ruben > Thanks - that worked. The other option I was thinking about was to use > different values for 'initialSeconds' in the intervalTrigger, for each > Project. Any pros/cons to either method? > > thanks > Michael > > On Jul 25, 10:25 pm, Ruben Willems <[email protected]> wrote: > > Hi > > > > place all projects in the same queue, and give each a different queue > > priority. > > This should solve it. > > > > with kind regards > > Ruben Willems > > > > On Sat, Jul 25, 2009 at 11:57 PM, Michael <[email protected]> wrote: > > > > > Hi, > > > I have 4 different projects setup in CCNet which use an > > > IntervalTrigger to check subversion for changes every 600 seconds. > > > This trigger works fine for each project individually - but when all 4 > > > projects trigger at the same time, it becomes a race condition to > > > access subversion. The first project will check for updates, but the > > > remaining three will generate an error: > > > > > Error Message: > > > ThoughtWorks.CruiseControl.Core.CruiseControlException: Source control > > > operation failed: svn: Write-lock stolen in 'E:\Builds\HW_Dept\trunk > > > \src' . Process command: svn cleanup E:\Builds\HW_Dept\trunk > > > > > I see with the ScheduleTrigger there is a random offset parameter to > > > prevent this exact thing when all projects are scheduled for the same > > > time - but there doesn't seem to be a way to stagger projects for an > > > intervalTrigger. > > > > > Basically, this becomes an issue if I restart the CCNet service, as > > > all projects will reset and schedule the next intervalTrigger build > > > time to be the exact same time (equivalent to a scheduleTrigger) > > > > > Any suggestions? > > > > > thanks > > > Michael >
