Can I delete error log files in apache. Will it cause an issue?

Question

Possible Duplicate:
Logrotate configuration for httpd (CentOS)

can I delete error log files in apache. Any issue? Because it eats my server space

Another "possible duplicate" that is not even close to being the same question. We need to crack down on this. Some answers provide partial coverage, but really this is nothing at all to do with logrotation and it may not even be implemented. — mckenzm, Jan 25 '23 at 23:09

score 16 · Answer 1 · edited Jan 13 '19 at 22:12

16

If you must clear the log file, do

cp /dev/null /var/log/file

or

echo > /var/log/file

That truncates the file without closing any open file handles.

Edit: Using logrotate to deal with the files is the better long-term solution. Truncating the file should be an emergency measure.

edited Jan 13 '19 at 22:12

kasperd

30,455
17
76
124

answered Oct 11 '12 at 17:57

Rob Bos

854
7
8

5

I like doing :>/var/log/file. It's quick, and looks like a bird. – Tom O'Connor Oct 12 '12 at 18:05
Thx, it helped me :) – Soren May 13 '20 at 19:13

score 3 · Answer 2 · answered Oct 11 '12 at 17:40

I would rather recommend going for logrotate in stead of deleting the old log files straight away. You might need them later for debugging something or to find out the pattern when system was running fine, for comparison purpose.

There is a reason that backup and recovery are so sought after. However if this is not any production server or something important, I wouldn't worry about deleting them.

score 1 · Answer 3 · answered Oct 11 '12 at 08:10

1

If you are not going to use them you can delete it. I would recomend you to delete all logs but not the current one, and if you can I should use log rotate.

answered Oct 11 '12 at 08:10

Hola Soy Edu Feliz Navidad

183
3
13

score 0 · Answer 4 · edited Apr 13 '17 at 12:37

0

If in use, deleting the log will make apache reload to re-open the just deleted file. What you can try is indeed logrotate, or try to empty the file. I think you'll need to SIGUSR1 apache though.

This link may help.

Edit: Just tested:

# > /var/log/apache2/access.log
# tail -F /var/log/apache2/access.log
.. Still prints lines ....
# rm /var/log/apache2/access.log ; touch /var/log/apache2/access.log
# tail -F /var/log/apache2/access.log
.. displays nothing ....
# kill -USR1 $(pidof apache2)
# tail -F /var/log/apache2/access.log
.. prints lines again ....

Edit2: thanks @voretaq for further info about signals on apache. SIGUSR1 is indeed less violent than SIGHUP.

edited Apr 13 '17 at 12:37

Community

1

answered Oct 11 '12 at 17:15

aif

381
1
8

2

A graceful restart (`SIGUSR1`) would generally be preferable to a "drop all connections and restart RIGHT FREAKIN' NOW!" (`SIGHUP`). – voretaq7 Oct 12 '12 at 17:11
1

Thanks @voretaq7 I didn't find out those information on the doc. Editing answer. – aif Oct 12 '12 at 17:34

Can I delete error log files in apache. Will it cause an issue?

4 Answers4