timja
commented
5 years ago An interesting and frustrating situation but I'm afraid I don't any good answers or suggestions for you.
Let me if I can summarize your situation. The agent log shows that the agent is connected. The master log shows that the agent has connected. But, the master UI is not showing the agent as connected and the system doesn't know it is connected to schedule builds on it. Does that describe it correctly?
If that's the case, it sounds like the connection is being made fine, but somehow it's not being added into all the required places inside the master. It's hard to guess what might be happening there.
I suggest combing through the master logs for anything relating to node, computer, agent, or slave to see if you can find any errors. You could try using the script console to poke around Jenkins.getNodes() and anything else related to see if you can find anything interesting.
Please report back with any results.
Hi Jenkins team/community
we started to observe that our cloud agents sometimes (roughly fail rate is
10%0.5% of launches) fail to come up online within first attempt. Here it a timeline and logs of such case:~13:40 agent was created by cloud plugin (VMware in this case, but same problem was observed on kubernetes cloud also, so i suggest cloud type is not important)
Here is agent log
here is part of server log related to same agent
but agent did not show up online at 13:40 as you can see on load statistic screenshot (0 executors available)
I waited 5 minutes and restarted jenkins windows service on agent VM at 13:46,
here is agent log
Server log
and after that agent connected and come online and started job execution , (see load statistic screenshot above)
We upgraded to 2.164.3 from 2.164.1 last week and noticed this issue but we are not 100% sure it wasn't there before upgrade.
Any Ideas or direction to troubleshoot further are appreciated
Thanks
Originally reported by iceiceice, imported from: agent connected via TCP but executor does not show up online sometimes