-
Bug
-
Resolution: Unresolved
-
Minor
-
Jenkins Version: 2.60.3
Amazon EC2 plugins: 1.36
Idle termination time: -5
I am seeing cases where Jenkins slaves are being stopped two or three times when using the ec2 plugin. I was able to go to the AWS console and confirm the slave was successfully stopped after the first request.
Update: We changed our local branch to only go through the logic if the agent isOnline() which seemed to resolve the issue.
From /var/log/jenkins/jenkins.log:
Aug 26, 2017 8:02:05 AM hudson.plugins.ec2.EC2RetentionStrategy internalCheck INFO: Idle timeout of jenkins-slave-name (i-xxxxxxxxxxxxxxxx) after 24 idle minutes, with 3 minutes remaining in billing period Aug 26, 2017 8:02:05 AM hudson.plugins.ec2.EC2AbstractSlave idleTimeout INFO: EC2 instance idle time expired: i-xxxxxxxxxxxxxxxx Aug 26, 2017 8:02:06 AM hudson.plugins.ec2.EC2AbstractSlave stop INFO: EC2 instance stop request sent for i-xxxxxxxxxxxxxxxx Aug 26, 2017 8:03:59 AM hudson.plugins.ec2.EC2RetentionStrategy internalCheck INFO: Idle timeout of jenkins-slave-name (i-xxxxxxxxxxxxxxxx) after 26 idle minutes, with 2 minutes remaining in billing period Aug 26, 2017 8:03:59 AM hudson.plugins.ec2.EC2AbstractSlave idleTimeout INFO: EC2 instance idle time expired: i-xxxxxxxxxxxxxxxx Aug 26, 2017 8:04:00 AM hudson.plugins.ec2.EC2AbstractSlave stop INFO: EC2 instance stop request sent for i-xxxxxxxxxxxxxxxx
From AWS console:
State transition reason: User initiated (2017-08-26 08:02:05 GMT) State transition reason message: Client.UserInitiatedShutdown: User initiated shutdown
I'm having the same issue but in my case any build taking over 30 min is broken because the node is shut down no matter what... Any ideas?