[Gluster-devel] uss.t in master doing bad things to our regression test VM's
Justin Clift
justin at gluster.org
Wed Feb 18 17:12:49 UTC 2015
Hi Vijaikumar,
As part of investigating what is going wrong with our VM's in Rackspace,
I created several new VM's (11 of them) and started a full regression
test run on them.
They're all hitting a major problem with uss.t. Part of it does a "cat"
on /dev/urandom... which is taking several hours at 100% of a cpu. :(
Here is output from "ps -ef f" on one of them:
root 12094 1287 0 13:23 ? S 0:00 \_ /bin/bash /opt/qa/regression.sh
root 12101 12094 0 13:23 ? S 0:00 \_ /bin/bash ./run-tests.sh
root 12116 12101 0 13:23 ? S 0:01 \_ /usr/bin/perl /usr/bin/prove -rf --timer ./tests
root 382 12116 0 14:13 ? S 0:00 \_ /bin/bash ./tests/basic/uss.t
root 1713 382 0 14:14 ? S 0:00 \_ /bin/bash ./tests/basic/uss.t
root 1714 1713 96 14:14 ? R 166:31 \_ cat /dev/urandom
root 1715 1713 2 14:14 ? S 5:04 \_ tr -dc a-zA-Z
root 1716 1713 9 14:14 ? S 16:31 \_ fold -w 8
And from top:
top - 17:09:19 up 3:50, 1 user, load average: 1.04, 1.03, 1.00
Tasks: 240 total, 3 running, 237 sleeping, 0 stopped, 0 zombie
Cpu0 : 4.3%us, 95.7%sy, 0.0%ni, 0.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Cpu1 : 8.1%us, 15.9%sy, 0.0%ni, 76.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 1916672k total, 1119544k used, 797128k free, 114976k buffers
Swap: 0k total, 0k used, 0k free, 427032k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1714 root 20 0 98.6m 620 504 R 96.0 0.0 169:00.94 cat
137 root 20 0 36100 1396 1140 S 15.9 0.1 37:01.55 plymouthd
1716 root 20 0 98.6m 712 616 S 10.0 0.0 16:46.55 fold
1715 root 20 0 98.6m 636 540 S 2.7 0.0 5:08.95 tr
9 root 20 0 0 0 0 S 0.3 0.0 0:00.59 ksoftirqd/1
1 root 20 0 19232 1128 860 S 0.0 0.1 0:00.93 init
2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd
Your name is on the commit which added the code, but that was months ago.
No idea why it's suddenly being a problem. Do you have any idea?
I am going to shut down all of these new test VM's except one, which I can
give you (or anyone) access to, if that would help find and fix the problem.
Btw, this is pretty important. ;)
+ Justin
--
GlusterFS - http://www.gluster.org
An open source, distributed file system scaling to several
petabytes, and handling thousands of clients.
My personal twitter: twitter.com/realjustinclift
More information about the Gluster-devel
mailing list