Procfs-related core dump deadlock
Greetings --
We've been experiencing some strange situations on our production machines, which look identical to the problems described in
<http://lists.insecure.org/linux-kernel/2001/Dec/1539.html>.
A follow-up to that message
(<http://lists.insecure.org/linux-kernel/2001/Dec/1604.html>) mentions that the problem is likely to be due to a deadlock associated with the coredump process.
The first message has some quoted text claiming that the problem went away in 2.4.10, but the poster is seeing it in 2.4.14, and we're still seeing it in 2.4.17-aa.
Does anybody have any idea of where to start looking, or what could have changed in the 2.4.10 -> 2.4.14 time frame that would have re-introduced this behavior?
We're currently kludging around the problem with 'ulimit -c 0' in relevant environments, but it would be nice to have a real fix.
john.
Signature · 2 lines
The only skills I have the patience to learn are those that have no real application in life. -- Calvin