sDovecot LMTP high CPU after migration from Dedicated Server to Virtual machine

imrans

• June 09, 2020 08:32

Dear all, I have just migrated few account from old centos 6 dedicated Server to Centos 7 Virtual machine . I am facing issue with high cpu usage due to Dovecot/lmtp multiple processes using 100% cpu . I have cheched the logs and queue and found below: - [root@srv1 ~]# ps ax | grep dovecot 4820 ? Ss 0:00 /usr/sbin/dovecot -F -c /etc/dovecot/dovecot.conf 4822 ? S 0:00 dovecot/pop3-login 4823 ? S 0:00 dovecot/imap-login 4824 ? S 0:00 dovecot/anvil 4825 ? S 0:00 dovecot/log 4826 ? S 0:00 dovecot/config 4827 ? S 0:00 dovecot/stats 4828 ? S 0:00 dovecot/pop3-login 4829 ? S 0:00 dovecot/imap-login 4830 ? S 0:00 dovecot/auth 8969 ? R 7:50 dovecot/lmtp 9477 ? R 6:55 dovecot/lmtp 9647 ? R 4:55 dovecot/lmtp 9778 ? R 2:55 dovecot/lmtp 9896 ? S 0:00 dovecot/lmtp Top command shows this:- PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 9477 apytpcom 20 0 119292 55792 7332 R 100.0 0.6 7:32.65 lmtp 9647 apytcom 20 0 94140 35568 7344 R 100.0 0.4 5:33.30 lmtp 9778 apytcom 20 0 94140 30412 7204 R 100.0 0.3 3:32.75 lmtp 8969 apytcom 20 0 119292 60836 7356 R 97.3 0.7 8:25.64 lmtp dovecot services status shows this :- Jun 09 13:24:02 srv1.vde3tel.com dovecot[4825]: lmtp(9896): Disconnect from local: Client has quit the connection (state=READY) Jun 09 13:25:01 srv1.vde3tel.com dovecot[4825]: lmtp(9896): Connect from local Jun 09 13:25:01 srv1.vde3tel.com dovecot[4825]: lmtp(9896): Disconnect from local: Client has quit the connection (state=READY) Jun 09 13:26:01 srv1.vde3tel.com dovecot[4825]: lmtp(9896): Connect from local Jun 09 13:26:01 srv1.vde3tel.com dovecot[4825]: lmtp(9896): Disconnect from local: Client has quit the connection (state=READY) Hint: Some lines were ellipsized, use -l to show in full. the above is current time . in Cpanel queue manager shows lots of messages like this

Sender	Size	apytcom@srv1.domain.com	1jie7h-0002Ri-KR	apytcom@srv1.domain.com	850 bytes	Queued

can somebody help me in this to fix this Dovecot/LMTP high CPU usage . Just to give some more information that "apytcom" is one domain hosted on this server and srv1.domain.com is the hostname of the virutal machine where cpanel is installed. Thank you in advance Imran about 500 messages

• 

  imrans

  • June 09, 2020 18:50

  Some more infomration for same issue, I have checked the dovecot log and found this very often Error: Mailbox list index was marked as fsck'd also found below error in log :- : Error: lmtp-server: conn unix:pid=1951,uid=47 [5]: rcpt apytcom@srv1.domain.com: Corrupted transaction log file /home/apytcom/mail/dovecot.list.index.log seq 3:: Error: lmtp-server: conn unix:pid=1951,uid=47 [5]: rcpt apytcom@srv1.domain.com: Corrupted transaction log file /home/apytcom/mail/dovecot.list.index.log seq 3: I have deleted the full queue and still these emails are being coming into queue . Please advise Thank you

  0
• 

  imrans

  • June 09, 2020 19:00

  Now I found one more error from log as below:- Jun 09 18:58:02 lmtp(apytcom@srv1.domain.com)<3293>: Debug: lmtp-server: conn unix:pid=3313,uid=47 [2]: rcpt apytcom@srv1.domain.com: Added userdb setting: plugin/userdb_quota_status_overquota=552 5.2.2 Mailbox is full / Blocks limit exceeded / Inode limit exceeded so all these issue and the major issue is LMTP high CPU all the time. Sorry for adding up all the information just to get to some solution from anyone in the community. Thank you

  0
• 

  cPanelLauren

  • June 11, 2020 16:52

  Deleting the mail queue won't resolve an issue where new mails are inbound into the queue. To clarify all of these emails inbound are indicating that the user's mailbox is full? Is the user sending mail out either legitimately or not? Resolving the actual problem here would be the key not the notification messages since they're automatically created.

  0
• 

  imrans

  • June 11, 2020 18:20

  Hi Lauren, Thank you for your Response. after reading your suggestion I moved forward and tried to find the main issue, and found our that there was a cron job which was running and all notification was being generated out of it. I have deleted that cron job and restarted the server. Now all is ok and no load nothing. Thank you for the help . you can resolve this thread .

  0

Please sign in to leave a comment.