server running out of memory often

[Grumpy]

The server is managed (not having ssh access)

If you have managed service, tell THEM to figure out what the exact problem is instead of just making a vague claim. I don't get why we're here.

There's ton of things they can do from measuring slow queries, slow executions, profiling, etc... You don't have ssh, you have no access, you lack the options to figure out as well as they can.

[Nevo]

Grumpy, on 22 Jan 2013 - 19:39, said:

If you have managed service, tell THEM to figure out what the exact problem is instead of just making a vague claim. I don't get why we're here.

There's ton of things they can do from measuring slow queries, slow executions, profiling, etc... You don't have ssh, you have no access, you lack the options to figure out as well as they can.

Even unmanaged would be able to tell you that... but clearly, the hosting this person has isn't professional.

[SG Staff]

Can you setup a Cron job? When I moved to a dedicated server, I set up a cron job to empty my memory out every hour. We used to have the out of memory issue constantly, but I haven't had it since. As a matter of fact, I don't even think I have gotten into using my Swap since I added this little script.

Cron Job set to run every hour every day. And this command as it looks.

sync; echo 3 > /proc/sys/vm/drop_caches

[Grumpy]

Can you setup a Cron job? When I moved to a dedicated server, I set up a cron job to empty my memory out every hour. We used to have the out of memory issue constantly, but I haven't had it since. As a matter of fact, I don't even think I have gotten into using my Swap since I added this little script.

Cron Job set to run every hour every day. And this command as it looks.

sync; echo 3 > /proc/sys/vm/drop_caches

This is a bad advice...
Dropping caches won't help anything. Additionally, following such advice should only result in worse performance. Caches are automatically dropped by the OS if the memory is needed by things like application.

If that cron job honestly helped you (rather than just coincidence), it would suggest that there's a bug with your kernel.

[Dmacleo]

wasn't that cron used to work around a cpanel mod_fcgid issue that cropped up only in pretty specific situations?

[Pross22]

IMO dropping caches isnt a good idea, defeats the idea of a cache in the first place and probably adds to server io as stuff has to be re-cached. The reason your running out of memory is nothing to do with the caches.

[SG Staff]

Well, alright then. Consider me corrected. :)

[mark2012]

After the upgrade to 3.4.2 everything seems to back to normal:
System Uptime 4 Days, 19 Hours and 22 Minutes

:flowers:

PS: Still get to optimize my datebase, will post at job section next week.

:tongue: