Article ID: 127042, created on Sep 29, 2015, last review on Sep 29, 2015

  • Applies to:
  • Plesk Automation 11.5

Symptoms

PA Control Panel perfomance is extremely slow. The average load of the system is too high:

# uptime
09:51:09 up 550 days, 12:34,  3 users,  load average: 3.22, 2.48, 2.70

The processes sw-engine-fpm: pool plesk consumes a lot of CPU time:

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
27798 psaadm    20   0  204m  20m 7556 S 84.4  0.1   0:25.73 sw-engine-fpm
28053 root      20   0  225m  52m  17m R 21.9  0.2   0:07.43 sw-engine
 5366 mysql     20   0  338m  56m 5188 S 13.0  0.2 140:07.77 mysqld
27788 psaadm    20   0  240m  54m 6916 S  4.6  0.2   0:13.88 sw-engine-fpm

A lot of records like proc /var/named/run-root/proc proc rw,relatime 0 0 in /etc/mtab:

[root ~]#cat /proc/mounts |grep named
...
proc /var/named/run-root/proc proc rw 0 0
vzfs /var/named/run-root/var/run/dbus vzfs rw 0 0
proc /var/named/run-root/proc proc rw 0 0
vzfs /var/named/run-root/var/run/dbus vzfs rw 0 0
proc /var/named/run-root/proc proc rw 0 0
vzfs /var/named/run-root/var/run/dbus vzfs rw 0 0`
...

or

...
proc /var/named/run-root/proc proc rw,relatime 0 0
proc /var/named/run-root/proc proc rw,relatime 0 0
proc /var/named/run-root/proc proc rw,relatime 0 0
proc /var/named/run-root/proc proc rw,relatime 0 0
proc /var/named/run-root/proc proc rw,relatime 0 0
...

There are can be thousands of them:

# cat /etc/mtab |wc -l
9978

Cause

Issue is caused by an sw-engine-fpm process that reads the content of /proc/mounts (/etc/mtab usually a symlink to /proc/mounts) which has lots of duplicated lines.

They are generated by named that fails to start and does not unmount properly even if named is stopped. The reason of the failed named in /var/log/messages:

...
Apr 15 12:26:02 ppa named: zone 0.0.127.IN-ADDR.ARPA/IN: loaded serial 20010622
Apr 15 12:26:02 ppa named: zone 1.168.192.in-addr.arpa/IN: loaded serial 1354556645
Apr 15 12:26:02 ppa named: zone example.com/IN: has no NS records
Apr 15 12:26:02 ppa named: _default/example.com/IN: bad zone
...

Resolution

  1. Fix broken DNS zone
  2. Restart named service:

    # service named stop && service named start
    
  3. Remove excessive lines from /etc.mtab:

    # for i in {0..9960};do  umount /var/named/run-root/proc;echo $i;done
    

Search Words

slow management node

mounts

mtab

proc

sw-engine-fpm

Медленно работает ssh консоль управляющей ноды

slow work

/var/named/run-root/proc

Медленно работает управляющая нода

много записей "proc /var/named/run-root/proc" в /etc/mtab

33a70544d00d562bbc5b17762c4ed2b3 caea8340e2d186a540518d08602aa065 e0aff7830fa22f92062ee4db78133079

Email subscription for changes to this article
Save as PDF