php-fpm (5.4.13) takes up all cpu resources

[email protected] (Free Proxy)
Newsgroups php.general
Message-ID <CAKM9q905HLDM=PQvX7d++aOCyph3VOBPykDLR4By=qH8mwPxBQ@mail.gmail.com>
Hi All,

I have a strange problem with php-fpm, I checkd lots of document, but I
can't fix it.
I'm using a Freebsd 91 server to host my website. Nginx and php-fpm 5.4.13
and mysql are using.
The hardware of my server is
12Core CPUs,
32G memory
6 HDD used as a Raidz2 array to host website
2 SSD used as a mirror array to host Database

After I upgrade from 5.4.11 o 5.4.13, I got below problem.

Usually, my system is in below status:(top)

last pid: 52977;  load averages:  4.70,  4.66,
5.55
301 processes: 2 running, 299 sleeping
CPU:  6.5% user,  0.0% nice,  6.3% system,  0.9% interrupt, 86.3% idle
Mem: 7029M Active, 11G Inact, 12G Wired, 1076M Cache, 399M Free
ARC: 8412M Total, 3190M MFU, 1763M MRU, 624K Anon, 209M Header, 3249M Other
Swap: 36G Total, 145M Used, 36G Free
  PID USERNAME    THR PRI NICE   SIZE    RES STATE   C   TIME   WCPU COMMAND
52728 www           1  25    0   150M 35564K accept 13   0:01  5.86% php-fpm
52735 www           1  25    0   150M 35416K accept 16   0:01  5.76% php-fpm
52740 www           1  25    0   150M 37668K accept 14   0:01  5.66% php-fpm
52739 www           1  24    0   150M 36424K accept 12   0:01  5.66% php-fpm
52708 www           1  24    0   150M 36324K accept 10   0:01  5.66% php-fpm
52724 www           1  24    0   150M 37740K select 16   0:01  5.57% php-fpm
52714 www           1  25    0   150M 36952K accept 21   0:01  5.57% php-fpm
52719 www           1  25    0   150M 37252K accept 23   0:01  5.57% php-fpm
52717 www           1  25    0   150M 39004K accept 12   0:01  5.57% php-fpm
52734 www           1  25    0   150M 37720K accept 18   0:01  5.57% php-fpm
52726 www           1  24    0   150M 37308K accept 23   0:01  5.47% php-fpm
52737 www           1  25    0   150M 37832K accept 14   0:01  5.37% php-fpm
52721 www           1  25    0   150M 36548K accept 22   0:01  5.37% php-fpm
52710 www           1  25    0   154M 38792K accept 16   0:01  5.37% php-fpm
52723 www           1  25    0   150M 35768K accept 10   0:01  5.37% php-fpm

But sometimes, I found a php-fpm will take up about 25%+ cpu resource, then
the cpu usage of php-fpm will get increased one by one. Then php-fpm will
take up all CPU resources.
Like below:

last pid: 54212;  load averages:  75.83,  75.55,
75.73
up 3+10:47:39  10:27:53
301 processes: 6 running, 295 sleeping
CPU:  6.8% user,  0.0% nice,  95.2% system,  0.7% interrupt, 0.0% idle
Mem: 7021M Active, 11G Inact, 12G Wired, 1358M Cache, 561M Free
ARC: 8525M Total, 3438M MFU, 1619M MRU, 1866K Anon, 213M Header, 3253M Other
Swap: 36G Total, 145M Used, 36G Free
  PID USERNAME    THR PRI NICE   SIZE    RES STATE   C   TIME   WCPU COMMAND
53939 www           1  31    0   150M 33096K accept 22   0:01  28.59%
php-fpm
53941 www           1  31    0   154M 38144K accept 20   0:01  28.40%
php-fpm
53938 www           1  30    0   150M 32816K accept  3   0:01  27.57%
php-fpm
53677 www           1  52    0   150M 28440K accept  2   0:04  26.49%
php-fpm
53948 www           1  27    0   150M 28964K accept 19   0:01  24.69%
php-fpm
53947 www           1  26    0   150M 29272K accept  8   0:01  24.05%
php-fpm
53951 www           1  25    0   150M 31208K accept 17   0:01  23.86%
php-fpm
53953 www           1  24    0   150M 35300K accept 15   0:00  23.86%
php-fpm
53950 www           1  25    0   150M 28268K accept  2   0:00  23.76%
php-fpm
53954 www           1  24    0   150M 35476K accept 19   0:00  23.76%
php-fpm
53955 www           1  24    0   150M 31292K accept 21   0:00  23.76%
php-fpm
53956 www           1  24    0   150M 34064K accept 19   0:00  23.66%
php-fpm
53962 www           1  24    0   150M 33224K CPU9    9   0:00  23.66%
php-fpm
53960 www           1  24    0   150M 33548K CPU17  17   0:00  23.66%
php-fpm


Below is the php-fpm.slow.

[01-Apr-2013 16:46:52] WARNING: [pool www] child 96623, script
'/web/www/forum.php' (request: "GET /forum.php") executing too slow
(30.661218 sec), logging
[01-Apr-2013 16:46:52] WARNING: [pool www] child 96620, script
'/web/www/forum.php' (request: "GET /forum.php") executing too slow
(31.778921 sec), logging
[01-Apr-2013 16:46:52] NOTICE: child 96764 stopped for tracing
[01-Apr-2013 16:46:52] NOTICE: about to trace 96764
[01-Apr-2013 16:46:52] NOTICE: finished trace of 96764
[01-Apr-2013 16:46:52] NOTICE: child 96763 stopped for tracing
[01-Apr-2013 16:46:52] NOTICE: about to trace 96763
[01-Apr-2013 16:46:52] ERROR: failed to ptrace(PT_IO) pid 96763: Bad
address (14)
[01-Apr-2013 16:46:52] NOTICE: finished trace of 96763
[01-Apr-2013 16:46:52] NOTICE: child 96762 stopped for tracing
[01-Apr-2013 16:46:52] NOTICE: about to trace 96762
[01-Apr-2013 16:46:52] NOTICE: finished trace of 96762
[01-Apr-2013 16:46:52] NOTICE: child 96761 stopped for tracing
[01-Apr-2013 16:46:52] NOTICE: about to trace 96761

The only way to fix this issue is to restart php-fpm.


At first, I though it's a mysql error, but I monitorred mysql process, and
I found that there is no slowlog. I tried to reinstall php, but this issue
still exist.

Can anyone help on this issue?
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.