# Nginx causes cpu burst problem when it is received USR1 signal by logrotate script **URL:** \ **Category:** NGINX\ **Created:** [October 18, 2010, 9:55am UTC](https://www.ruby-forum.com/t/nginx-causes-cpu-burst-problem-when-it-is-received-usr1-signal-by-logrotate-script/197648 "2010-10-18T09:55:35Z")\ **Posts on this page:** 5\ **Page:** 1
**Author:** ![okkezSS](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/o/85f322/32.png) [@okkezSS](https://www.ruby-forum.com/u/okkezSS)\ **Post date:** [October 18, 2010, 9:55am UTC](https://www.ruby-forum.com/t/nginx-causes-cpu-burst-problem-when-it-is-received-usr1-signal-by-logrotate-script/197648/1 "2010-10-18T09:55:35Z")
Hi, I’m using nginx/0.7.67 on the freebsd-6.4 $ nginx -V nginx version: nginx/0.7.67 built by gcc 2.95.3 20010315 (release) configure arguments: --prefix=/usr/local –conf-path=/usr/local/conf/nginx/nginx.conf –pid-path=/usr/local/var/run/nginx.pid –lock-path=/usr/local/var/run/nginx.lock –http-client-body-temp-path=/usr/local/var/run/nginx/client\_body\_temp –http-proxy-temp-path=/usr/local/var/run/nginx/proxy\_temp –error-log-path=/usr/local/logs/nginx/error –http-log-path=/usr/local/logs/nginx/access --with-select\_module –with-http\_perl\_module --with-debug --with-http\_stub\_status\_module I’m using the following script as root user for logrotation: ```auto #!/bin/sh mv /usr/local/logs/nginx/error /usr/local/logs/nginx/error.`date "+%Y%m%d"` mv /usr/local/logs/nginx/access /usr/local/logs/nginx/access.`date "+%Y%m%d"` kill -USR1 `cat /usr/local/var/run/nginx.pid` find /usr/local/logs/nginx/ -ctime +30 -exec rm {} \; ``` Log rotation succeed. but, Occasionally the above script causes master process’ cpu burst until next logrotate time. top command outputs following: [code] $ top last pid: 91548; load averages: 1.00, 1.02, 1.01 101 processes: 5 running, 78 sleeping, 18 waiting CPU: 7.8% user, 0.0% nice, 43.4% system, 0.0% interrupt, 48.8% idle Mem: 91M Active, 1617M Inact, 181M Wired, 68M Cache, 213M Buf, 19M Free Swap: 4096M Total, 4K Used, 4096M Free PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND 41661 root 1 113 0 27172K 7716K CPU1 1 486:03 99.02% nginx 11 root 1 171 52 0K 16K RUN 0 ??? 94.97% idle: cpu0 10 root 1 171 52 0K 16K RUN 1 ??? 1.56% idle: cpu1 41673 nginx 1 96 0 27480K 8040K select 0 11:27 0.05% nginx 12 root 1 -44 -163 0K 16K WAIT 0 20.6H 0.00% swi1: net 22 root 1 -68 0 0K 16K - ``` 0 872:49 0.00% em0 taskq ``` 24 root 1 -68 -187 0K 16K WAIT 0 638:23 0.00% irq16: em0 uhci0+ 13 root 1 -32 -151 0K 16K WAIT 0 476:52 0.00% swi4: clock sio 43 root 1 20 0 0K 16K syncer 0 144:55 0.00% syncer 41 root 1 171 52 0K 16K pgzero 0 130:24 0.00% pagezero 46 root 1 -32 0 0K 16K - ``` 0 39:09 0.00% schedcpu ``` 3 root 1 -8 0 0K 16K - 0 35:15 0.00% g\_up 4 root 1 -8 0 0K 16K - 0 31:01 0.00% g\_down 2 root 1 -8 0 0K 16K - 0 25:47 0.00% g\_event 6635 root 1 96 0 14024K 2308K select 0 25:16 0.00% sshd 640 root 1 96 0 7576K 1692K select 0 20:49 0.00% ntpd 41672 nginx 1 96 0 27484K 8048K select 0 11:56 0.00% nginx 41692 nginx 1 96 0 27484K 8044K select 0 11:55 0.00% nginx 41666 nginx 1 96 0 27480K 8044K select 0 11:51 0.00% nginx 41682 nginx 1 96 0 27480K 8040K select 0 11:50 0.00% nginx 41691 nginx 1 96 0 27488K 8048K select 0 11:50 0.00% nginx 41663 nginx 1 96 0 27476K 8040K select 0 11:50 0.00% nginx 41664 nginx 1 96 0 27476K 8040K select 0 11:48 0.00% nginx 41676 nginx 1 96 0 27480K 8040K RUN 0 11:46 0.00% nginx 41686 nginx 1 96 0 27484K 8044K select 0 11:45 0.00% nginx 41671 nginx 1 96 0 27476K 8040K select 0 11:43 0.00% nginx 41687 nginx 1 96 0 27480K 8040K select 0 11:43 0.00% nginx 41662 nginx 1 96 0 27476K 8040K select 0 11:40 0.00% nginx 41685 nginx 1 96 0 27480K 8040K select 0 11:40 0.00% nginx 41675 nginx 1 96 0 27480K 8040K select 0 11:37 0.00% nginx 41681 nginx 1 96 0 27480K 8040K select 0 11:37 0.00% nginx 41680 nginx 1 96 0 27476K 8036K select 0 11:37 0.00% nginx 41674 nginx 1 96 0 27476K 8036K select 0 11:36 0.00% nginx 41690 nginx 1 96 0 27484K 8044K select 0 11:34 0.00% nginx 41688 nginx 1 96 0 27480K 8040K select 0 11:33 0.00% nginx 41670 nginx 1 96 0 27484K 8048K select 0 11:33 0.00% nginx 41689 nginx 1 96 0 27484K 8044K select 0 11:33 0.00% nginx [code] normally top command outputs following: [code] $ top last pid: 43789; load averages: 0.00, 0.00, 0.00 108 processes: 3 running, 87 sleeping, 18 waiting CPU: 0.0% user, 0.0% nice, 0.4% system, 0.0% interrupt, 99.6% idle Mem: 102M Active, 1604M Inact, 181M Wired, 73M Cache, 213M Buf, 16M Free Swap: 4096M Total, 4K Used, 4096M Free PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND 10 root 1 171 52 0K 16K CPU1 1 ??? 99.02% idle: cpu1 11 root 1 171 52 0K 16K RUN 0 ??? 97.61% idle: cpu0 60947 root 1 96 0 27172K 7712K select 0 24.0H 0.00% nginx 12 root 1 -44 -163 0K 16K WAIT 0 20.8H 0.00% swi1: net 22 root 1 -68 0 0K 16K - ``` 0 884:49 0.00% em0 taskq ``` 24 root 1 -68 -187 0K 16K WAIT 0 644:02 0.00% irq16: em0 uhci0+ 13 root 1 -32 -151 0K 16K WAIT 0 481:46 0.00% swi4: clock sio 43 root 1 20 0 0K 16K syncer 0 144:23 0.00% syncer 41 root 1 171 52 0K 16K pgzero 1 128:31 0.00% pagezero 46 root 1 -16 0 0K 16K - ``` 0 40:48 0.00% schedcpu ``` 3 root 1 -8 0 0K 16K - 0 35:22 0.00% g\_up 4 root 1 -8 0 0K 16K - 0 31:20 0.00% g\_down 2 root 1 -8 0 0K 16K - 0 26:14 0.00% g\_event 6652 root 1 96 0 14024K 2332K select 0 24:07 0.00% sshd 640 root 1 96 0 7576K 1692K select 0 20:52 0.00% ntpd 60951 nginx 1 96 0 27476K 8032K select 0 13:17 0.00% nginx 60968 nginx 1 96 0 27484K 8040K select 0 13:14 0.00% nginx 60972 nginx 1 96 0 27488K 8044K select 0 13:12 0.00% nginx 60956 nginx 1 96 0 27580K 8140K select 0 13:10 0.00% nginx 60954 nginx 1 96 0 27480K 8040K select 0 13:09 0.00% nginx 60974 nginx 1 96 0 27484K 8040K select 0 13:05 0.00% nginx 60969 nginx 1 96 0 27480K 8036K select 0 12:59 0.00% nginx 60978 nginx 1 96 0 27484K 8040K select 0 12:59 0.00% nginx 60967 nginx 1 96 0 27480K 8036K select 0 12:59 0.00% nginx 60977 nginx 1 96 0 27484K 8040K select 0 12:59 0.00% nginx 60949 nginx 1 96 0 27480K 8036K select 0 12:57 0.00% nginx 60948 nginx 1 96 0 27476K 8032K select 0 12:57 0.00% nginx 60975 nginx 1 96 0 27480K 8036K select 0 12:57 0.00% nginx 60953 nginx 1 96 0 27476K 8036K select 0 12:55 0.00% nginx 60965 nginx 1 96 0 27484K 8040K select 0 12:54 0.00% nginx 60961 nginx 1 96 0 27480K 8036K select 0 12:54 0.00% nginx 60958 nginx 1 96 0 27476K 8036K select 0 12:53 0.00% nginx 60970 nginx 1 96 0 27480K 8036K select 0 12:52 0.00% nginx 60964 nginx 1 96 0 27480K 8036K select 0 12:52 0.00% nginx 60966 nginx 1 96 0 27480K 8036K select 0 12:52 0.00% nginx 60976 nginx 1 96 0 27484K 8040K select 0 12:52 0.00% nginx 60971 nginx 1 96 0 27480K 8036K select 0 12:50 0.00% nginx [code] Am I doing something wrong? Regards, Posted at Nginx Forum: > **[nginx causes cpu burst problem when it is received USR1 signal by logrotate...](https://forum.nginx.org/read.php?2,141795,141795#msg-141795)** > > Hi, I'm using nginx/0.7.67 on the freebsd-6.4 $ nginx -V nginx version: nginx/0.7.67 built by gcc 2.95.3 20010315 (release) configure arguments: --prefix=/usr/local --conf-path=/usr/local/conf/nginx/nginx.conf --pid-path=/usr/local/var/run/nginx.pid... ---
**Author:** ![hakutoitoi](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/ba9def/32.png) [@hakutoitoi](https://www.ruby-forum.com/u/hakutoitoi)\ **Post date:** [October 18, 2010, 11:06am UTC](https://www.ruby-forum.com/t/nginx-causes-cpu-burst-problem-when-it-is-received-usr1-signal-by-logrotate-script/197648/2 "2010-10-18T11:06:52Z")
On Mon, Oct 18, 2010 at 03:54:12AM -0400, hakutoitoi wrote: > –http-client-body-temp-path=/usr/local/var/run/nginx/client\_body\_temp > –http-proxy-temp-path=/usr/local/var/run/nginx/proxy\_temp > –error-log-path=/usr/local/logs/nginx/error > –http-log-path=/usr/local/logs/nginx/access --with-select\_module > –with-http\_perl\_module --with-debug --with-http\_stub\_status\_module > 41673 nginx 1 96 0 27480K 8040K select > 0 11:27 0.05% nginx Why do you use the poor select method ? FreeBSD has very good kqueue method. This CPU burst may be caused by nginx bug in select method handling. – Igor S. > **[Igor Sysoev](http://sysoev.ru/en/)** ---
**Author:** ![hakutoitoi](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/ba9def/32.png) [@hakutoitoi](https://www.ruby-forum.com/u/hakutoitoi)\ **Post date:** [October 18, 2010, 5:31pm UTC](https://www.ruby-forum.com/t/nginx-causes-cpu-burst-problem-when-it-is-received-usr1-signal-by-logrotate-script/197648/3 "2010-10-18T17:31:35Z")
Igor, Thank you for your advice. I wanted to use kqueue and have configured using kqueue settings. /usr/local/conf/nginx/nginx.conf ```auto events { worker_connections 1024; use kqueue; multi_accept off; } ``` error log outputs using kqueue event method. /usr/local/logs/nginx/error ```auto 2010/10/18 20:48:32 [notice] 77055#0: using the "kqueue" event method 2010/10/18 20:48:32 [notice] 77055#0: nginx/0.7.67 ``` but, I also use Embedded Perl module. ```auto ./configure --with-http_perl_module ``` configure with --with-http\_perl\_module ```auto $ top 69650 nginx 1 4 0 2044K 1540K kqread 0 0:00 0.00% nginx ``` configure without --with-http\_perl\_module ```auto $ top 77061 nginx 1 96 0 3828K 2124K select 0 0:00 0.00% nginx ``` as long as I use the Embedded Perl module, I can’t fix this problem? Posted at Nginx Forum: > **[Re: nginx causes cpu burst problem when it is received USR1 signal by...](https://forum.nginx.org/read.php?2,141795,141922#msg-141922)** > > Hi, I'm using nginx/0.7.67 on the freebsd-6.4 $ nginx -V nginx version: nginx/0.7.67 built by gcc 2.95.3 20010315 (release) configure arguments: --prefix=/usr/local --conf-path=/usr/local/conf/nginx/nginx.conf --pid-path=/usr/local/var/run/nginx.pid... ---
**Author:** ![hakutoitoi](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/ba9def/32.png) [@hakutoitoi](https://www.ruby-forum.com/u/hakutoitoi)\ **Post date:** [October 18, 2010, 6:15pm UTC](https://www.ruby-forum.com/t/nginx-causes-cpu-burst-problem-when-it-is-received-usr1-signal-by-logrotate-script/197648/4 "2010-10-18T18:15:17Z")
On Mon, Oct 18, 2010 at 11:29:46AM -0400, hakutoitoi wrote: > } > > [code] > configure without --with-http\_perl\_module > > ```auto > $ top > 77061 nginx 1 96 0 3828K 2124K select 0 0:00 0.00% nginx > ``` > > as long as I use the Embedded Perl module, I can’t fix this problem? Probably, you mixed up: configure with --with-http\_perl\_module kqueue configure without --with-http\_perl\_module select It seems that you use perl code that works with remote servers or databases and this select is called by some perl module. – Igor S. > **[Igor Sysoev](http://sysoev.ru/en/)** ---
**Author:** ![hakutoitoi](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/ba9def/32.png) [@hakutoitoi](https://www.ruby-forum.com/u/hakutoitoi)\ **Post date:** [October 19, 2010, 4:36am UTC](https://www.ruby-forum.com/t/nginx-causes-cpu-burst-problem-when-it-is-received-usr1-signal-by-logrotate-script/197648/5 "2010-10-19T04:36:02Z")
sorry I mixed up. configure with --with-http\_perl\_module kqueue configure without --with-http\_perl\_module select I tried simple config file(no use perl syntax) ```auto $ cat /usr/local/conf/nginx/nginx.conf user nginx users; error_log logs/nginx/error notice; worker_rlimit_nofile 8192; worker_processes 32; pid /usr/local/var/run/nginx.pid; events { worker_connections 1024; use kqueue; multi_accept off; } #include conf.d/*.conf; ``` configure with --with-http\_perl\_module ```auto $ top PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND 1551 root 1 4 0 3760K 664K select 1 3:31 0.00% syslogd 1605 root 1 8 0 3700K 432K nanslp 1 1:45 0.00% cron 1600 root 1 4 0 14020K 952K select 1 0:38 0.00% sshd 52642 root 1 4 0 24928K 2996K sbwait 0 0:00 0.00% sshd 5953 root 1 4 0 3592K 1916K select 1 0:00 0.00% nginx 5959 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5979 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5966 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5963 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5972 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5975 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5977 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5954 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5958 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5988 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5976 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5986 nginx 1 96 0 3828K 2140K select 0 0:00 0.00% nginx 5974 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5981 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5955 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5965 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5957 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5962 nginx 1 96 0 3828K 2132K select 1 0:00 0.00% nginx 5985 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5961 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5973 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5964 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5968 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5967 nginx 1 96 0 3828K 2132K select 0 0:00 0.00% nginx 5978 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5983 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5987 nginx 1 96 0 3828K 2140K select 0 0:00 0.00% nginx 5982 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5970 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5980 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5984 nginx 1 96 0 3828K 2136K select 0 0:00 0.00% nginx 5971 nginx 1 4 0 3828K 2136K select 1 0:00 0.00% nginx ``` configure without --with-http\_perl\_module ```auto $ top PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND 1551 root 1 96 0 3760K 664K select 0 3:31 0.00% syslogd 1605 root 1 8 0 3700K 432K nanslp 0 1:45 0.00% cron 1600 root 1 4 0 14020K 952K select 1 0:38 0.00% sshd 99672 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 52642 root 1 4 0 24928K 2996K sbwait 0 0:00 0.00% sshd 99701 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99695 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99678 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99694 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99707 nginx 1 4 0 2020K 1508K kqread 0 0:00 0.00% nginx 99674 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99692 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99680 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99700 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99687 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99677 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99675 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99688 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99690 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99676 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99702 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99697 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99681 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99673 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99685 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99682 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99699 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99704 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99689 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99691 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99683 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99706 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99693 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99679 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99698 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx 99671 root 1 20 0 1796K 1296K pause 0 0:00 0.00% nginx 99703 nginx 1 4 0 2020K 1512K kqread 0 0:00 0.00% nginx ``` If I built nginx with --with-http\_perl\_module, it seems that nginx uses select method instead of kqueue. Posted at Nginx Forum: > **[Re: nginx causes cpu burst problem when it is received USR1 signal by...](https://forum.nginx.org/read.php?2,141795,142174)** > > Hi, I'm using nginx/0.7.67 on the freebsd-6.4 $ nginx -V nginx version: nginx/0.7.67 built by gcc 2.95.3 20010315 (release) configure arguments: --prefix=/usr/local --conf-path=/usr/local/conf/nginx/nginx.conf --pid-path=/usr/local/var/run/nginx.pid...