I am experiencing a perplexing problem with an openSuse 11.4 install on which I run Moodle 2.0.3+ (Build: 20110511).
Once every 24 hours, every day, I must reboot my server because we lose our ability to access Moodle via the web interface - even on the server itself. I don’t know if this is an Apache problem, a Moodle problem, or something else in the O/S. Can anyone recommend some steps to isolate the cause of this problem?
bkreps wrote:
> I am experiencing a perplexing problem with an openSuse 11.4 install on
> which I run Moodle 2.0.3+ (Build: 20110511).
> Once every 24 hours, every day, I must reboot my server because we lose
> our ability to access Moodle via the web interface - even on the server
> itself. I don’t know if this is an Apache problem, a Moodle problem, or
> something else in the O/S. Can anyone recommend some steps to isolate
> the cause of this problem?
You say that “we lose our ability to access Moodle via the web
interface”. Does that mean that (a) you can still access moodle via some
other interface and (b) you can still access other things via the web
interface (static HTML pages for example)?
This sounds like an interaction/configuration issue with moodle to me.
Does it use a database for example, and is cron backing up the database
every night perhaps?
Do you always lose access to moodle at the same time? What time? What
else is going on at that time?
You may have better luck asking in the moodle community.
Back to the basics. When you loose connectivity:
- If you have more than computer on an network, ping one of the other computers from the server and vica versa. If this does nor work, run
ifconfig -a
and investigate / post.
- See if you can ping another external site i.e.“www.google.com”
- Open a web browser and connect to a site.
You should by now have a clearer picture of where the problem lies.
Give me 24 hours and I’ll provide a reply.
The problem recurred this morning and I first determined that I had full network connectivity. I then issued ‘/etc/init.d/apache2 -status’ which returned the result of ‘Dead’. I then issued ‘/etc/init.d/apache2 -restart’. The response was ‘Syntax OK’, ‘Warning: found stale pidfile (unclean shutdown?)’, and ‘Starting httpd2 (prefork)’. Done.
The question that I must determine is why Apache is shutting down. My next step will be to look for the appropriate log.
I’m experiencing the same problem with apache, did you solve your problem ? I’m using Moodle 2.1.2+ (Build: 20111012) .
By googleing for “apache2 seg fault or similar nasty error detected in the parent process” i found a potential solution here :
It seems that a 3rd party module that kills httpd when rotating logs with ‘service httpd reload’ and Changing this to ‘restart’ in the logrotate httpd script, sorts the issue out.
So I replaced /etc/rotate.d/apache2 at the postrotate comand /etc/init.d/apache2 reload by /etc/init.d/apache2 restart.
I don’t know yet whether my problem is solved and I wonder if you or anyone found the cause and a solution.
bkreps, i have been having a similar problem. Apache just dies. After checking the error_log , i find the message;
[notice] seg fault or similar nasty error detected in the parent process
And googling does give hints that the problem could be due to logrotate for apache which happens daily around 4am.
What i have done so far is to ;
nano /etc/logrotate.d/apache2
replace /etc/init.d/apache2 reload with /usr/sbin/rcapache2 restart
Looks like rcapache2 handles the stale pidfile better.
Will let you know if this works or fails. I am testing this workaround now.
Regards,
Marco
p.s - On 11.4 64 bits.
Confirmed - The workaround works.
