Yes, the Jenkins user can ssh to the slave without a password. I have even verified that this is possible in the Jenkins Groovy console (I was able to run a shell script in that user's home directory, and that script successfully used ssh). The slave launching process seems to fail somewhere before actually running a script (in the "run a script to launch" configuration).
-- Todd From: [email protected] [mailto:[email protected]] On Behalf Of Chanda Unmack Sent: Tuesday, May 21, 2013 8:20 PM To: [email protected] Subject: Re: Unable to start ssh slave from OSX master I came to that conclusion in our environment, for many other reasons. I inherited the setup which had a master on osx, with slaves connected via ssh (linux & osx) as well as jnlp (windows) which is why I didn't feel that I could offer up any help. The only thing I wasn't clear on with your set up is whether or not the user you specify jenkins to ssh to the slave as can log into the slave without entering a password. chanda On Tue, May 21, 2013 at 5:05 PM, Todd Greer <[email protected]<mailto:[email protected]>> wrote: In case anyone comes across this in web searches, I've tried many different things, and now suspect that using an OSX master with ssh slaves may simply not be a workable configuration, for unknown reasons. I've seen plenty of reports that it does work well with a Linux master, so that is my intent, and my suggestion to anyone in my situation. -- Todd From: Todd Greer Sent: Friday, May 17, 2013 2:48 PM To: '[email protected]<mailto:[email protected]>' Subject: Unable to start ssh slave from OSX master I'm trying to use the ssh slave plugin (v.21) from OSX, and have been unable to get it to initiate the ssh connection. The node is configured with the username and password, using the default port (which is correct). From a manual connection, there is an entry for the slave node in ~/.ssh/known_hosts. Other types of slave nodes are fine. When I try to start the node, or to look at the node's logs, I just get a spinner. In Jenkins's logs, I see "Excess workload 1.0 detected. (planned capacity=0.0,Qlen=0.99999976,idle=0.0&0,total=0m,=0.5)". In system.log, I see "Attempting to reconnect <node-name>". Other than that, I see no signs of activity, including in the slave node's auth log. Can anyone provide me with guidance on where to look next? Thank you, Todd -- Todd Greer Principal Scientist, Affinegy, Inc. -- You received this message because you are subscribed to the Google Groups "Jenkins Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]<mailto:jenkinsci-users%[email protected]>. For more options, visit https://groups.google.com/groups/opt_out. -- You received this message because you are subscribed to the Google Groups "Jenkins Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]<mailto:[email protected]>. For more options, visit https://groups.google.com/groups/opt_out. -- You received this message because you are subscribed to the Google Groups "Jenkins Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
