(Please use 'reply' and not 'reply-all'... I don't need two copies of your replies, one in my inbox <G>)
As I mentioned before, instance cap management in the current ec2-plugin is a bit suspect, although I would have only expected it to fail in the other direction (too many instances, not too few). ----- Original Message ----- From: [email protected] To: [email protected] Cc: KEVIN FLEMING (BLOOMBERG/ 731 LEXIN) At: Jan 22 2013 12:53:05 Ok. Rebooted the box and it now works. I tried increasing the caps - did not work reload jenkins-ci - did not work reboot, so restart jenkins - worked Cap value must only being read at startup time then. Will have to find time to full test this, and get into the jenkins bug tracker. Anyway thank for the replies. On Tuesday, January 22, 2013 4:28:28 PM UTC, Kevin Fleming wrote: > > Yes, I have a global instance cap and also a cap on each slave template > (although per-template caps don't work properly in the current ec2-plugin > version... I have patches to fix that that I'll be submitting soon). I > believe I saw proper resumption of stopped instances even before I fixed > the instance cap handling though. > > You may be running into another bug in the ec2-plugin though, for which I > also have a fix but it requires Jenkins core changes. When you see a new > instance getting spawned unnecessarily, is the stopped instance *also* > resumed but left idle? > > ------------------------------ > *From:*[email protected] <javascript:>*To:* > [email protected] <javascript:>*Cc:*KEVIN FLEMING (BLOOMBERG/ > 731 LEXIN)*Date:* 1/22 11:25 > ------------------------------ > > Hi, thanks for the reply. > > In your settings do you have a cap on the number of instance you can > launch? > > Otherwise I am not sure what's wrong since it a very basic setup and > there's little in the logs to go on. > > On Tuesday, January 22, 2013 2:47:56 PM UTC, Kevin Fleming wrote: >> >> Yes, I use it heavily with my slave templates configured to 'stop' >> instead of 'terminate', and the plugin properly restarts the slaves when >> they are needed. >> >> ------------------------------ >> *From:*[email protected]*To:*[email protected]*Date:* >> 1/20 >> 8:57 >> ------------------------------ >> >> Hi, would log this as a bug but I can't login to jenkins jria (ldap >> error). >> >> The plugin seems to work fine apart from automatically waking up a slave >> which is stopped. >> >> Logs say it want to provision another node, but it should wake up the >> stopped node and use that. >> >> If I go to the node page and start the slave it comes up fine and the job >> runs. >> >> Jenkins - 1.480.2 LTS >> Amazon EC2 >> plugin<http://wiki.jenkins-ci.org/display/JENKINS/Amazon+EC2+Plugin> - >> 1.17 >> Master with no executors, set to no be used >> 1 slave and can only provision 1 aws node >> >> Anybody use this plugin, with stopped instead of terminate? >> >> Any help would be great. >> >
