[2989] in Release_7.7_team
Re: Linux update problems
daemon@ATHENA.MIT.EDU (andrew m. boardman)
Tue Oct 2 11:11:51 2001
Date: Tue, 2 Oct 2001 11:11:46 -0400
Message-Id: <200110021511.LAA14755@pothole.mit.edu>
From: "andrew m. boardman" <amb@MIT.EDU>
To: zacheiss@mit.edu
CC: release-team@mit.edu
In-reply-to: <200110020736.DAA11170@hodge-podge.mit.edu> (message from Garry
Zacheiss on Tue, 02 Oct 2001 03:36:51 -0400)
>we tracked down this problem to a modprobe of the usb-uhci kernel module
>that appeared to go into kernel wait and then never came back; this made
>a kudzu never exit and so update_ws never exited.
Can you point me at specific machines on which this happened? (A copy of
a relevant modules.conf would be lovely.) I have a few theories, but
they're all probably wrong because none of them explain why the machines
with the new modules.conf didn't hang on subsequent boots, too.
What was the success/failure rate you saw for cluster machines?