aborted_clients every 6 minutes [Case 81093]

jimlongo

Well-Known Member
Mar 20, 2008
242
20
68
Hi, still trying to clean this up. The number of aborted clients is still relatively high (>10%)

I've found that there is an aborted connection for 'root' every 6 minutes like so . . .

Code:
131021 16:14:00 [Warning] Aborted connection 1525 to db: 'mysql' user: 'root' host: 'localhost' (Unknown error)
131021 16:20:09 [Warning] Aborted connection 1750 to db: 'mysql' user: 'root' host: 'localhost' (Unknown error)
131021 16:26:20 [Warning] Aborted connection 1849 to db: 'mysql' user: 'root' host: 'localhost' (Unknown error)
My timeout in my.cnf is set to 600

What I'm wondering if you have any idea what could be happening with such regularity?

Of if there is a way I can track it down. I've tried watching mysql status, but it doesn't really give me any clue.
Is there anything else I can look at?

Thanks.
 

jimlongo

Well-Known Member
Mar 20, 2008
242
20
68
Thanks,

I increased the wait_timeout to 1200, the aborted_clients are still the same 6 minutes apart.


Code:
131022  0:58:53 [Warning] Aborted connection 31 to db: 'mysql' user: 'root' host: 'localhost' (Unknown error)
131022  1:05:04 [Warning] Aborted connection 56 to db: 'mysql' user: 'root' host: 'localhost' (Unknown error)
131022  1:11:11 [Warning] Aborted connection 80 to db: 'mysql' user: 'root' host: 'localhost' (Unknown error)
131022  1:17:14 [Warning] Aborted connection 111 to db: 'mysql' user: 'root' host: 'localhost' (Unknown error)
 

jimlongo

Well-Known Member
Mar 20, 2008
242
20
68
Thanks, I enabled all the logs and if I match up by timestamp the aborted_client message from the error log with the query from the general log it gives me this query as the cause.
Code:
22 Connect   [email protected] as anonymous on mysql
22 Query     set sql_mode='' /* DISABLE ANSI_QUOTES and normalize */
22 Query     SELECT user FROM mysql.user WHERE user LIKE 'cpses_%'
Can you identify what process is using that query?
 

jimlongo

Well-Known Member
Mar 20, 2008
242
20
68
From the cpanel error log there is this process which matches the timestamp of the aborted_client and query . . .

Code:
[2013-10-22 12:13:06 -0400] info [cpses_tool] cleaning up orphened cpses_ users: Id: TQ:TaskQueue:3222
I have no database user like cpses_%
Is there a way to avoid this aborted_client?
 

cPanelMichael

Administrator
Staff member
Apr 11, 2011
47,913
2,200
363
I recommend opening a support ticket so we can take a closer look:

Submit A Ticket

You can post the ticket number here so we can update this thread with the outcome.

Thank you.
 

jimlongo

Well-Known Member
Mar 20, 2008
242
20
68
Had to switch from MariaDB back to MySQL before support would look at this.

Support Request Id 4377179
 
Last edited:

jimlongo

Well-Known Member
Mar 20, 2008
242
20
68
According to this support request, "This is related to a bug that's already been filed. The case number is 81093".
 

cPanelMichael

Administrator
Staff member
Apr 11, 2011
47,913
2,200
363
It's been determined this is a bug where cpses_tool does not properly disconnect from MySQL. For reference, the case number is 81093. You can monitor the change log for this case number to determine when a resolution is pushed out:

cPanel Change Log

Thank you.
 

jimlongo

Well-Known Member
Mar 20, 2008
242
20
68
This is fixed in 11.40.17

Fixed case 81093: cpses_tool is too verbose and does not properly disconnect from sql servers.

Thank you.