High Load average on NFS Spool - v.2.1.15 & 2.2.13

[Albert E. Whale, CEH CHS CISA CISSP]

I have two servers pointing to an NFS mounted mail spool with dovecot.  
Since I recently switched from using Dovecot v1.X, I have been 
experiencing high CPU use with the two Dovecot servers. I am not certain 
why they are not well behaved.  Here is the configuration information.


This configuration is currently running at a load average of 17.

/usr/sbin/dovecot -n
# 2.1.15: /etc/dovecot/dovecot.conf
# OS: Linux 3.10.54-server-2.mga3 i686 Mageia 3
base_dir = /var/run/dovecot/
disable_plaintext_auth = no
lock_method = dotlock
login_greeting = SpamZapper Email ready.
login_log_format_elements = user=<%u> method=%m rip=%r lip=%l %c
mail_debug = yes
mail_fsync = always
mail_location = maildir:~/Maildir
mail_nfs_index = yes
mail_nfs_storage = yes
mail_privileged_group = mail
mbox_lock_timeout = 15 mins
mmap_disable = yes
passdb {
   driver = pam
}
protocols = imap pop3
service anvil {
   client_limit = 203
   process_limit = 1
}
service auth {
   user = root
}
service imap-login {
   chroot = login
   client_limit = 10
   user = dovecot
   vsz_limit = 256 M
}
service pop3-login {
   chroot = login
   client_limit = 20
   user = dovecot
}
ssl = no
syslog_facility = local2
userdb {
   driver = passwd
}
verbose_proctitle = yes
protocol pop3 {
   mail_max_userip_connections = 20
   pop3_logout_format = top=%t/%p, retr=%r/%b, del=%d/%m, size=%s
   pop3_uidl_format = %08Xv%08Xu
}
protocol lda {
   postmaster_address = postmaster at example.com
}
protocol imap {
   mail_max_userip_connections = 10
}

And the other server which is currently running with a load average of 10.

dovecot -n
# 2.2.13: /etc/dovecot/dovecot.conf
doveconf: Warning: service anvil { client_limit=100 } is lower than 
required under max. load (303)
# OS: Linux 4.1.15-server-2.mga5 x86_64 Mageia 4
base_dir = /var/run/dovecot/
disable_plaintext_auth = no
listen = *
lock_method = dotlock
login_greeting = SpamZapper Email ready.
login_log_format_elements = user=<%u> method=%m rip=%r lip=%l %c
mail_fsync = always
mail_location = maildir:~/Maildir
mail_nfs_index = yes
mail_nfs_storage = yes
mail_privileged_group = mail
mmap_disable = yes
passdb {
   driver = pam
}
protocols = imap pop3
service anvil {
   client_limit = 100
   process_limit = 1
}
service auth {
   user = root
}
service imap-login {
   chroot = login
   client_limit = 48
   user = dovecot
   vsz_limit = 256 M
}
service pop3-login {
   chroot = login
   client_limit = 50
   user = dovecot
}
ssl = no
syslog_facility = local2
userdb {
   driver = passwd
}
verbose_proctitle = yes
protocol pop3 {
   mail_max_userip_connections = 50
   pop3_logout_format = top=%t/%p, retr=%r/%b, del=%d/%m, size=%s
   pop3_uidl_format = %08Xv%08Xu
}
protocol lda {
   postmaster_address = postmaster at example.com
}
protocol imap {
   mail_max_userip_connections = 40
}

Can you please offer a suggestion on what is happening?  The timeouts to 
the imap server are frequent due to the load average.

Thank you.

-- 
-- 
--

Albert E. Whale, CEH CHS CISA CISSP
*President - Chief Info Security Officer*
IT Security, Inc. <http://www.IT-Security-inc.com> - A Service Disabled 
Veteran Owned Company - (*SDVOSB*)
*HUBZone Certified*
LinkedIn <https://www.linkedin.com/in/albertwhale> Profile


Phone: 412-515-3010 | Email: Albert.Whale at IT-Security-inc.com
Cell: 412-889-6870

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://dovecot.org/pipermail/dovecot/attachments/20181218/083a500a/attachment.html>