Apache log-rotate feature is not working

[mmgenius]

We have used the httpd errorlog rotate feature but it didn't work today, there was created a domain-error.log of 14GB.
(https://www.directadmin.com/features.php?id=1939)

directadmin.conf content:
rotate_httpd_error_log_meg=1024
rotate_httpd_error_log_truncate=1

Do we need to do anything else?

 

[factor]

there was created a domain-error.log of 14GB.

is very big you might need to do something manually.

 

[mmgenius]

is very big you might need to do something manually.

This is the reason why I want to put it on max 1GB using 'rotate_httpd_error_log_meg=1024'
(before I changed directadmin.conf I deleted the 14GB file manually and restarted httpd)

 

[factor]

Ok it was already set to rotate at a gig and it didnt rotate?

OR

You want it to rotate at a gig now because you didn't have it set before the 14 g log file?

If you have it set now it should rotate at a gig.

 

[mmgenius]

Ok it was already set to rotate at a gig and it didnt rotate?

It was already the second time, I change directadmin.conf after the first time. But some days later the problem was there again and he didn't rotate.

 

[factor]

Humm might be a bug then. Assuming the os and da is all updated?

 

[mmgenius]

indeed all updated.

 

[mmgenius]

again, log-file ran full without rotating

 

[mmgenius]

This is what I found in /var/log/directadmin/errortaskq.log

2022:01:11-16:03:02: removing old lock: ./data/admin/error_log_truncate.lock.lock (age: 3601 seconds) this caller='Apache::check_for_rotate_error_logs', old creation info: lock created Tue Jan 11 15:03:01 2022
2022:01:11-16:03:02: An error occurred while sending the mail
2022:01:11-16:03:02: An error occurred while sending the mail
2022:01:11-16:03:02: Apache::check_for_rotate_error_logs: couldn't open truncate log /var/log/httpd/domains/xxxxxxx.xxx.error.log.trunc for writing: No space left on device
2022:01:11-16:04:34: An error occurred while sending the mail
2022:01:11-16:04:34: An error occurred while sending the mail
2022:01:11-16:04:34: Apache::check_for_rotate_error_logs: couldn't open truncate log /var/log/httpd/domains/xxxxxxx.xxx.error.log.trunc for writing: No space left on device
2022:01:11-16:05:02: An error occurred while sending the mail
2022:01:11-16:05:03: An error occurred while sending the mail
2022:01:11-16:05:03: Apache::check_for_rotate_error_logs: couldn't open truncate log /var/log/httpd/domains/xxxxxxx.xxx.error.log.trunc for writing: No space left on device
2022:01:11-16:06:02: An error occurred while sending the mail
2022:01:11-16:06:02: An error occurred while sending the mail
2022:01:11-16:06:02: Apache::check_for_rotate_error_logs: couldn't open truncate log /var/l2022:01:11-16:32:02: service mysqld wasn't running, starting mysqld

 

[Driesp]

You probably need to add space to this partition.
I think it can't rotate if there is no space left.

 

[mmgenius]

That is the problem.
It is a drive of 15GB, the log_rotate is set to 1GB, but the file is growing to 14GB.
So there is space enough, but because the log_rotate doesn't work is gets full.

 

[Driesp]

I think you only have 2 options, or you resize the partition or you try to fix an issue with an application that writes 14GB of errorlog per day.
I would suggest the latter.

 

[mmgenius]

Yes, the thing that need to be done is fix the application.
But it is a share hosting system, so we have to ask our customer to look at it. And tomorrow there could be another script that have the problem.
DirectAdmin has a feature for it, so we like to use it, but for now it isn't working like it should work.

 

[Driesp]

Just disable php error logging for that client.

 

[wila]

Did you restart directadmin after making the change?

 

[mmgenius]

yes

 

[mmgenius]

And again he did not do a error_log rotate. He made again an error file of 14GB