Stefan,
I have some boxes that seem to reproduce this behaviour rather frequently.
This is great news since this bug is so hard to reproduce.
I just launched with the new testing kernels you provided (aki-9ab546f3
x86_64) on a server that has 4096meg cache (the bad behaving size).
Unfortunately I'm already seeing signs that this kernel has the CPU
reporting bug where CPU times are in the thousands of days. Usually
this means the box will be crashing in the next 1-2 days.. Will post
back when I confirm that the fork bug still exists in these kernels.
The CPUtime bug still exists. It was speculated that these were
unrelated.. This machine was just launched and within minutes it
displayed wrong cpu times.
process list snip from top:
26614 root 20 0 570m 448m 4972 S 0 2.9 0:39.41 ruby
26671 root 20 0 539m 419m 4948 S 0 2.7 0:35.79 ruby
28441 root 20 0 528m 409m 4212 S 0 2.7 24207187w ruby
26310 root 20 0 523m 405m 5040 S 3 2.6 0:51.46 ruby
29320 root 20 0 523m 405m 4948 S 0 2.6 91970855d ruby
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is a direct subscriber.
https://bugs.launchpad.net/bugs/708920
Title:
Strange 'fork/clone' blocking behavior under high cpu usage on EC2
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
