Remote agent was unresponsive and has gone offline

[Chris Moore]

Hi guys

I get the above error on one of my agents (Mac) at least once overnight.

I have 2 other agents (1 Mac and 1 PC) that don't receive this message.

I have a process that runs via launchctl, which calls a batch file to start the remote agent but that doesn't get executed, I believe because the agent bootstrap process is still running.

Could someone let me know how I should go about debugging the problem please? And which log files I should be looking at? Any settings should be the same for all remote agents as I don't think I've ever changed anything.

Cheers,

Chris.

[Chris Moore · Accepted solution]

Thank you for your help Peter in narrowing it down to being a Mac configuration issue rather than a Bamboo issue.

Turns out the Mac in question was scheduled to sleep every day at 11pm and then awaken again at 6am. I turned that off yesterday and all seems fine this morning so that must have been messing with the contact between the server and the Mac.

[Peter Van de Voorde]

Hi Chris,

This looks more like a configuration issue on your Mac, which I don't really have any experience with so I can't help you out there I'm sorry.

I would just say: take a look at differences in system settings between your two Macs (maybe something about sleep mode?).

Best regards,

Peter

[Chris Moore]

OK, thank you for your help Peter.

I should add that I have been working with Bamboo for around 6 months and this problem has only started happening in the last couple of weeks.

Regards,

Chris.

[Chris Moore]

Hi Peter

Thank you for your suggestions.

I tried implementing the auto start of Bamboo on the Mac but it doesn't seem to have worked. I added the plist and replaced the username with "XGS Mcbuild", ran the command and restarted but nothing happened.

What files should I be looking at for configuration details of the machines please and what settings should I be looking for?

Or is there an alternative method I could try? Basically, if I could shut down the terminal windows and agent bootstrap process associated with bamboo and then run my script to start the agent, I think it would be fine also. Basically a force re-run as that's what I have to do to fix it manually currently every morning.

Regards,

Chris.

[Chris Moore]

Hi Peter

Thank you for the information.

I will give the method described in the link a try.

Yes, there is a 2nd mac agent and also a PC agent that doesn't have these problems. I'm pretty sure the bamboo configurations on these would be identical as I was the one who set everything up and I didn't change any configurations.

Regards,

Chris.

[Peter Van de Voorde]

Hi Chris,

You'll need to refresh your agents lists to see changes in online/offline states.

The log is automatically refreshed but the page isn't.

The following page might provide you with some extra information:

https://confluence.atlassian.com/display/BAMKB/Configuring+Bamboo+to+start+automatically+on+startup+on+Mac+OSX

You also mentioned a second Mac agent that doesn't have these problems.

Are the configurations of these machines identical?

Best regards,

Peter

[Chris Moore]

Hi Peter

Thank you for your response.

I'm not entirely sure- when I look at it in the morning, I see at least one of the messages "Remote agent xxx was unresponsive and has gone offline" but when I look in my agents list, the agent is still listed as online, which seems contradictory.

What I do at that point is manually run a script that starts the remote agent:
cd /Users/xgsmcbuild/Desktop/Bamboo
java -jar bamboo-agent-5.2.jar http://fts0:8085/agentServer

I then get the following messages:

03-Apr-2014 08:49:17 Remote agent [10.10.0.70] marked as inactive. A new one came in place.

03-Apr-2014 08:49:17 Remote agent [10.10.0.70] came back after a period of inactivity.

This is on a mac so I have a launchd job that should run that script every minute via launchctl.

To be honest, I'm only just starting out with launchd and launchctl so there may potentially be an issue with that but it seems OK as if I manually quit out from the java agent bootstrap, the script does get run and the agent comes back as described above.

So I'm unsure of why the agent is becoming unresponsive and going offline. From what I've read, it seems to be something to do with a 'heartbeat' that gets sent from server to agent or the other way around so it must be something is wrong there. But I'm just unsure of what to check or how to debug it.

Regards,

Chris.

[Peter Van de Voorde]

Hi Chris,

Did you check if the agent did actually go offline? Or was it a network problem which caused Bamboo not to find your agent?

Does your agent restart automatically after a certain period of time?

Best regards,

Peter

[Manjunatha NS]

how to setup automatically restart the agents ? can you please explain