-
Bug
-
Resolution: Fixed
-
Major
-
None
-
a busy master :-)
If a master is too busy 1 to reply to the ping request with one minute, slaves run away en masse.
The symptoms are weird EOFExceptions during job executions, so jobs are terminated halfway.
Log
[04/07/10 13:31:24] Launching slave agent
$ /usr/bin/ssh 192.168.102.14 ~/common/slave.sh
<===[HUDSON REMOTING CAPACITY]===>channel started
Slave.jar version: 1.336 (private-11/28/2009 08:34-kohsuke)
This is a Unix slave
Copied maven-agent.jar
Copied maven-interceptor.jar
Copied maven2.1-interceptor.jar
Apr 7, 2010 1:32:59 PM com.youdevise.hudson.slavestatus.SlaveListener call
INFO: Slave-status listener starting
Apr 7, 2010 1:32:59 PM com.youdevise.hudson.slavestatus.SocketHTTPListener waitForConnection
INFO: Slave-status listener ready on port 3141
Ping failed. Terminating
[04/07/10 15:17:50] slave agent was terminated
java.io.EOFException
at java.io.ObjectInputStream$BlockDataInputStream.peekByte(ObjectInputStream.java:2498)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1273)
at java.io.ObjectInputStream.readObject(ObjectInputStream.java:348)
at hudson.remoting.Channel$ReaderThread.run(Channel.java:852)
- Some likely causes are: busy with scm polling or TCP window (or other buffering). This still needs investigation.