Running out of ram?

[webservice]

Hi,

I bought this system I while back. Installed it on a remote rented server (Fedora Core 5), no problem there. Works great, like it a lot!

Only problem is, this is a rather poor server, with only 512 MB RAM. After about 3-5 days of up-time, the server suddenly shuts down, no response from any service (http, ssh, ftp, pop3...).

Each time this happens, the people at ServerPronto.com say that: "there were no error on the console, so I restarted the server"...hmmm.

Anyway, I can see, that shortly before it crashes, the RAM is really pushing the limit, at around 96%-99% (mostly cached stuff).

So my question is: does the Zervex SCP cause this crash, because it needs more ram? Should I get a bigger server? Is 512 MB just too little?

Any advice much appreciated.


Thanks.

Rói Højgaard

[webservice]

Oops, I neglected to read the forum rules first. Sorry. If this post is considered unfit, don't hesitate to delete it, as I have now submitted a support request to http://support.zervex.com.


Thanks.

Rói

[Jeremy]

DSM is very light weight, so that's definitely not your problem.

What proc do you have? Running FC5 is more for the latest systems and i would never tell people to run it. Centos 4.4 is great for P4's down to a duron but for the older gen cpus centos 3.8 is better.

512mb is ok for a few sites, but nothing more... If you have a database MySQL is probably slowing down your system.

I would make sure your not running any extra services you don't need.

I would ask them to swap out your ram, Normally if your box kernel panic's it spits out an error or there not really looking... and i think that's the case ServerPronto has a crappy reputation for there servers and support.

Maybe your box is dieing off from swap thats what you really need to look out for.

Install cacti to monitor your servers resources and see what was going on before it crashes/locks up.

can you ping it?

[webservice]

Hi Jeremy,

thanks for your reply.

The remote server is fairly new (½ year old), using an AMD Sempron 2400+ CPU. And it actually has Fedora Core 4, not 5.

There are several services running, for example httpd, mysqld, courier, named, mailman, webmin and proftpd, to name a few. This was supposed to be a multi purpose webserver.

No, no response from any service when down, including ping.

I'm not exactly a Linux guru yet, but to me, there doesn't seem to be any swap space at all!? Viewing PhpSysInfo on my local box, it says "Disk Swap 0%" in the "Memory Usage" part, but no such thing on my rented server. Maybe they do this setup on purpose, as they take extra $$ to reboot a server? Dunno...

Attached is a screen dump from PhpSysInfo, on my newly rebooted remote server.

[Jeremy]

That's an OK proc i wouldn't call it a server class cpu....

I would see about adding swap, that's the heart and sole of Linux, other then the kernel.
http://www.redhat.com/docs/manuals/l...ap-adding.html
RedHat is the base for FC.

If you like... i would ask them to do a reinstall and setup the filesystem correct..... / /home /etc /usr /tmp and swap... But i think they did it so you have extra space.

There's not much i can tell you, i don't have access to the box but if you want to hire someone look at rack911.com I've worked with them a lot.

I'm not a fan of FC anyways

[webservice]

Ok, Jeremy, thanks a lot for your most useful input.

I feel much more confident about the Zervex SCP now.


Regards

Rói Højgaard

[Jeremy]

Not that it was help but yea