Description of problem: I noticed the 'YumBackend.py get-updates' process had wedged when a yum install told me it was blocking on a 38 day old process. I killed the YumBackend process, and a new one was respawned that noticed there is a distribution upgrade available. This then proceeded to notify me (54 times at once so far, and no end in sight), causing the Xorg process to consume close to 100% CPU and the system therefore unresponsive. I'd rather not reboot to clear this notification backlog as I have a large amount of work state that would need to be recreated if I did. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. Not update to Fedora 11 2. Wedge the yumBackend.py process that checks for updates 3. Wait 5 or 6 weeks 4. Kill the yumBackend.py process 5. Spend several hours clicking on dozens of simultaneous notifications while trying to preserve work Actual results: Dozens of simultaneous notifications Expected results: One notification Additional info:
Difficult to reproduce, but I think I can fudge something.
Hmm, I've been trying for a little while, but I can't reproduce this on F11 or rawhide. Does this still happen in F11?
No idea, I'm not running F11 yet - and even if I were it's a bit difficult to trigger in normal use :)
Right, I've reviewed the codebase we use in F11 and rawhide, and there's no way we can send multiple requests in a short amount of time. I'll close this one for now, but if you can reproduce with F11, I would be interested. Note., F11 seems to get itself wedged a lot less than F10!