High CPU issue with streaming on
Jammy
jammy.linux at gmail.com
Wed Mar 14 03:26:47 CET 2012
#0. We got varnish 3.0.2 running on box 2.6.32-220.4.1.el6.x86_64, with streaming on.
#1. the process
top - 22:19:21 up 14 days, 22:28, 2 users, load average: 10.86, 10.87, 19.51
Tasks: 1 total, 0 running, 1 sleeping, 0 stopped, 0 zombie
Cpu(s): 64.9%us, 0.5%sy, 0.0%ni, 0.0%id, 0.0%wa, 0.0%hi, 0.2%si, 34.5%st
Mem: 7646264k total, 6948476k used, 697788k free, 109384k buffers
Swap: 4194296k total, 7636k used, 4186660k free, 3502984k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
24551 nobody 20 0 4241m 40m 18m S 143.2 0.5 50:07.07 varnishd
#2. the threads
top - 22:19:35 up 14 days, 22:29, 2 users, load average: 11.50, 11.02, 19.42
Tasks: 637 total, 11 running, 626 sleeping, 0 stopped, 0 zombie
Cpu(s): 4.5%us, 0.8%sy, 0.0%ni, 92.3%id, 0.3%wa, 0.0%hi, 0.3%si, 1.9%st
Mem: 7646264k total, 6948856k used, 697408k free, 109396k buffers
Swap: 4194296k total, 7636k used, 4186660k free, 3503000k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
24757 nobody 20 0 4241m 40m 18m R 22.9 0.5 8:10.68 varnishd
24759 nobody 20 0 4241m 40m 18m R 21.5 0.5 7:59.45 varnishd
24958 nobody 20 0 4241m 40m 18m R 21.5 0.5 8:59.62 varnishd
24761 nobody 20 0 4241m 40m 18m R 16.2 0.5 5:26.20 varnishd
24954 nobody 20 0 4241m 40m 18m R 16.2 0.5 7:46.43 varnishd
24952 nobody 20 0 4241m 40m 18m R 14.8 0.5 4:20.41 varnishd
24960 nobody 20 0 4241m 40m 18m R 14.8 0.5 9:16.54 varnishd
#3. the stacktraces
Thread 206 (Thread 0x7f5047012700 (LWP 24757)):
#0 0x000000360df3e606 in __memcpy_ssse3_back () from /lib64/libc.so.6
#1 0x0000003610206ac7 in ?? () from /usr/lib64/varnish/libvgz.so
#2 0x00000036102072c6 in inflate () from /usr/lib64/varnish/libvgz.so
#3 0x00000000004233aa in VGZ_Gunzip ()
#4 0x0000000000423ce1 in VGZ_WrwGunzip ()
#5 0x000000000042f702 in RES_StreamPoll ()
#6 0x0000000000423607 in ?? ()
#7 0x0000000000421e1a in FetchBody ()
#8 0x0000000000416ffd in ?? ()
#9 0x000000000041785c in CNT_Session ()
#10 0x000000000042f508 in ?? ()
#11 0x000000000042e6eb in ?? ()
#12 0x000000360e2077f1 in start_thread () from /lib64/libpthread.so.0
#13 0x000000360dee592d in clone () from /lib64/libc.so.6
Thread 204 (Thread 0x7f5045c10700 (LWP 24759)):
#0 0x0000003610206cb4 in inflate () from /usr/lib64/varnish/libvgz.so
#1 0x00000000004233aa in VGZ_Gunzip ()
#2 0x0000000000423ce1 in VGZ_WrwGunzip ()
#3 0x000000000042f702 in RES_StreamPoll ()
#4 0x0000000000423607 in ?? ()
#5 0x0000000000421e1a in FetchBody ()
#6 0x0000000000416ffd in ?? ()
#7 0x000000000041785c in CNT_Session ()
#8 0x000000000042f508 in ?? ()
#9 0x000000000042e6eb in ?? ()
#10 0x000000360e2077f1 in start_thread () from /lib64/libpthread.so.0
#11 0x000000360dee592d in clone () from /lib64/libc.so.6
Most of the threads eating lots of CPUs were all with the similar stacktraces, and never finished...
I doubt there must be a infinite-loop or something similar.
@list, any suggestions?
Thanks
----------------------------------
Best wishes,
Jammy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.varnish-cache.org/lists/pipermail/varnish-dev/attachments/20120314/dcfb6fb3/attachment.html>
More information about the varnish-dev
mailing list