1

UPDATE: I noticed a similar issue, perhaps related. But I changed the MPM module to MPM event as noted in this GDB BT:

#0  0x00007fc74ea90df4 in read () from target:/lib64/libpthread.so.0
#1  0x000055df44ca24cb in ap_mpm_podx_check ()
#2  0x00007fc74e1f4f57 in ?? () from target:/etc/httpd/modules/mod_mpm_event.so
#3  0x00007fc74e1f528a in ?? () from target:/etc/httpd/modules/mod_mpm_event.so
#4  0x00007fc74e1f531b in ?? () from target:/etc/httpd/modules/mod_mpm_event.so
#5  0x00007fc74e1f6017 in ?? () from target:/etc/httpd/modules/mod_mpm_event.so
#6  0x000055df44c76bde in ap_run_mpm ()
#7  0x000055df44c6f2d3 in main ()

This started happening after the upgrade to Fedora 29. First MySQL starts using a lot of CPU, but as in the other thread, Apache/httpd then starts to really consume CPU.

top - 16:58:34 up 1 day,  5:14,  2 users,  load average: 22.06, 21.55, 21.13
Tasks: 273 total,   1 running, 272 sleeping,   0 stopped,   0 zombie
%Cpu(s): 12.5 us,  1.0 sy,  8.0 ni, 76.8 id,  1.6 wa,  0.1 hi,  0.0 si,  0.0 st
MiB Mem :   7968.8 total,    217.8 free,   1812.1 used,   5938.9 buff/cache
MiB Swap:   8000.0 total,   7547.9 free,    452.1 used.   5843.4 avail Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
12985 apache    20   0 2339912  17908  10748 S  99.3   0.2   3:42.00 /usr/sbin/httpd

I tried to run a gdb and bt on the process ID and got a lot of Missing separate debuginfo for target but it ends with:

0x00007f3e4a31cdf4 in read () from target:/lib64/libpthread.so.0
(gdb) bt
#0  0x00007f3e4a31cdf4 in read () from target:/lib64/libpthread.so.0
#1  0x00005640ecf6c4cb in ap_mpm_podx_check ()
#2  0x00007f3e49a128de in ?? () from target:/etc/httpd/modules/mod_mpm_worker.so
#3  0x00007f3e49a12c89 in ?? () from target:/etc/httpd/modules/mod_mpm_worker.so
#4  0x00007f3e49a12d1b in ?? () from target:/etc/httpd/modules/mod_mpm_worker.so
#5  0x00007f3e49a13b7f in ?? () from target:/etc/httpd/modules/mod_mpm_worker.so
#6  0x00005640ecf40bde in ap_run_mpm ()
#7  0x00005640ecf392d3 in main ()
(gdb) quit
A debugging session is active.

A strace on the process number just hangs at strace: Process 12985 attached read(8,

And Squirrelmail/webmail is timing out with 504 Gateway Timeout. The only errors in the Apache error logs don't seem to be relevant:

[Fri Jan 04 16:07:01.341027 2019] [proxy_fcgi:error] [pid 12985:tid 139904902215424] [client 201.141.38.249:51260] AH01071: Got error 'Primary script unknown\n'
<standard input>:13515: warning [p 113, 7.8i]: can't break line
<standard input>:13668: warning [p 115, 0.2i]: can't break line
<standard input>:686: warning [p 7, 4.5i, div `3tbd0,0', 0.2i]: can't break line
<standard input>:692: warning [p 7, 4.5i, div `3tbd0,3', 0.0i]: can't break line
<standard input>:694: warning [p 7, 4.5i, div `3tbd0,4', 0.0i]: can't break line
<standard input>:696: warning [p 7, 4.5i, div `3tbd0,5', 0.0i]: can't break line
<standard input>:700: warning [p 7, 4.5i, div `3tbd0,7', 0.0i]: can't break line
warning: file `<standard input>', around line 691:
  table wider than line width
<standard input>:1459: warning [p 11, 5.0i]: can't break line
<standard input>:40: warning [p 1, 2.5i]: can't break line
<standard input>:85: warning: can't find special character `u4F60'
<standard input>:85: warning: can't find special character `u597D'
<standard input>:347: warning [p 4, 7.8i, div `3tbd1,0', 0.0i]: can't break line
<standard input>:360: warning [p 4, 7.8i, div `3tbd4,0', 0.0i]: can't break line
<standard input>:380: warning [p 4, 7.8i, div `3tbd8,0', 0.0i]: can't break line
<standard input>:424: warning [p 5, 0.2i, div `3tbd1,0', 0.0i]: can't break line
<standard input>:424: warning [p 5, 0.2i, div `3tbd1,0', 0.3i]: can't break line
<standard input>:485: warning [p 5, 0.2i, div `3tbd14,0', 0.0i]: can't break line
<standard input>:1379: warning [p 13, 12.0i, div `3tbd10,1', 0.0i]: can't break line
<standard input>:1384: warning [p 13, 12.0i, div `3tbd11,0', 0.0i]: can't break line
<standard input>:1903: warning [p 15, 20.8i]: can't break line
[Fri Jan 04 16:59:13.393915 2019] [proxy_fcgi:error] [pid 12981:tid 139904482776832] [client 119.185.41.131:54918] AH01071: Got error 'Primary script unknown\n', 

This is not a particular busy server and the access_logs aren't filling up nor being filled so I don't think this is a DOS.

2
  • Since MySQL also starts using a lot of CPU resources, I would expect the problem does not lie with Apache but with Fedora. Also because it started since your upgrade. Are there any patch updates out? Perhaps move over to CentOS instead of running Fedora which does not sound suited for running servers.
    – Tommiie
    Jan 8, 2019 at 9:24
  • I decided to create a bug at bugzilla.redhat.com/show_bug.cgi?id=1665692, as it may or may not be httpd but it sure seems like, it. I added some strrace, lsof and gdb results.
    – RobbieTheK
    Jan 14, 2019 at 3:35

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.