Fetcher daemon crashes every few days

[jberryman]

I'm not sure what other relevant information to give. But monit seems to be having to relaunch the fetcher daemon between every 1-5 days. I suppose it's crashing, and I'm not sure if this would show up in any logs anywhere.

[look]

Sorry you're having problems with this. Rest assured you are not alone.

I have basically stopped using the daemon because for whatever reason it won't stay running (in my case the problem was more that the process would be running but wouldn't do anything).

I switched to using cron with a lockfile to execute the fetcher periodically which works pretty well; however sometimes it still hangs.

I added SystemTimer to the mix to help fix these problems. I think in some/most cases it's due to a problem in the Ruby interpreter which can prevent timeouts from working on system calls (like opening a network connection). You can read up on SystemTimer for more on that: http://ph7spot.com/musings/system-timer

[subimage]

Look - I'm having issues with cron/lockfile "locking up" and reporting that my fetcher process is still running when it obviously isn't.

Is this similar to the problems you encountered? Could you post an example of your cron/lockfile/systemtimer configuration for us to look at?

Appreciate it.

[look]

I've only had problems with it locking up when it really is still running (unless I kill it manually, in which case the lockfile is not deleted).

As for my configuration, it's basically what is in the readme file.

Wish I could be more helpful...