• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Extremely slow - mchk, spammng, drwebmng

awitt

New Pleskian
hi,

I have a server with Quad-Core Xeon processor and hardware RAID with 4 hard disk SAS 15k rpm.
The system has 1 domains 10 domains aliases and 150 mailboxes.
Adding another mail account takes 5-15 minutes.
Checking the system through 'mchk' works extremely slowly (about 2-4h). Server load 0.25, cpu usage ~ 25%, the disk IO ~ 1%.

What is wrong with your scripts? Whether they can run faster?
I have 6 servers with Plesk Control Panel, and on each is the same.

System: (debian, fedora, centos), CPU 2-4 Core AMD or Xeon, RAID / 15k rpm disk, Plesk Version: 9.2.3, 9.3.0


Has anything made by Parallels will work properly?
 
It is really very strange and absolutely non-typical. Investigation directly on server is required.
Therefore for effectively researching reasons of this problem I suggest you contact support team. It is very difficult to find a reasons of such issues with forum's discussions. We can daily discuss about possible reasons without success.
 
Strange and absolutely non-typical? Hmm...
Please do the same test in their LAB.
I wasted nearly three hours to perform this test.

Tested on a clean system in the following configuration using Plesk developer license:

PowerEdge SC1435
Amd Quad Core Opteron 2350, Raid1 2x 15krpm SAS
1 domain, 10 domain aliases, 150 mailboxes (no data)
--------------------------------------------------
Results below:

- adding 1 mailbox (/usr/local/psa/bin/mail)
0m10.230s

-adding 150 mailboxes (/usr/local/psa/bin/mail)
19m58.116s

- remove 1 mailbox (/usr/local/psa/bin/mail)
0m12.267s

- remove 150 mailboxes (/usr/local/psa/bin/mail)
26m32.868s

- running mchk (/usr/local/psa/admin/sbin/mchk)
49m16.493s

- backup client account (/usr/local/psa/bin/pleskbackup)
0m4.436s - backup file size 593KB

-restore client account (/usr/local/psa/bin/pleskrestore)
48m32.842s - backup file size 593KB

-delete client account (/usr/local/psa/bin/client)
26m42.690s


Is removing the client through web panel must end with a message: "MySQL has gone away"?
Whether the addition of e-mail accounts must occupy 20 minutes?
Whether to delete an account must be of almost 27 minutes?
Verification of empty mailboxes (no data) through MCHK lasts almost 50 minutes!
Is it normal that scripts running so slow?

What hardware is recommended for use by Parallels Plesk Control Panel?
 
Do you really think that each Plesk in whole world works so extremely slowly?
Did you tried logs investigation, for example?
Post here output of following commands at least:

# mysql -uadmin -p`cat /etc/psa/.psa.shadow`
mysql> show variables like '%max_con%';

and value of wait_timeout from /etc/my.cnf
 
Probably all of the servers with Plesk Panel have the same problem (in this configuration) :p

Mysql error is not a big problem for me.
The problem is the duration of the command mchk, adding a mailbox, restore the backup.
On a clean system, which does nothing time 50 minutes (running mchk) seems too long.

below the data that you asked:

+--------------------+-------+
| Variable_name | Value |
+--------------------+-------+
| max_connect_errors | 100 |
| max_connections | 512 |
+--------------------+-------+
2 rows in set (0.00 sec)

/etc/mysql/my.cnf
wait_timeout = 300

sorry for my English... :)
 
Default Plesk installation:

mysql> show variables like '%max_con%';
+--------------------+-------+
| Variable_name | Value |
+--------------------+-------+
| max_connect_errors | 10 |
| max_connections | 100 |
+--------------------+-------+
2 rows in set (0.00 sec)

mysql> show variables like '%wait_time%';
+--------------------------+-------+
| Variable_name | Value |
+--------------------------+-------+
| innodb_lock_wait_timeout | 50 |
| table_lock_wait_timeout | 50 |
| wait_timeout | 28800 |
+--------------------------+-------+
3 rows in set (0.00 sec)

Do you have any errors in /var/log/sw-cp-server/error_log?
Try to set wait_timeout to default value 28800
 
Will changing these values to the default you will see that it will run faster?
Let's leave the problem with the error "MySQL has gone away". It's not important to me at the moment.
Log is clean (/var/log/sw-cp-server/error_log). Contains only the entry: (log.c.75) server started.

I've set the default value in mysql. I run a short test ... (add 150 mailboxes 1client, 1domain, 10domain aliases)...
 
Back
Top