Hi,

I got weird problem with my dedicated (posted in the dedicated forum, but got no response - probably wrong forum? ;)):

Today, I was unable to connect to http, ssh, ftp, pop... in other words i was unable to connect to anything.
On the other hand, I was able to ping it. Though, pings were very bad (the were average 160ms, with around 36ms normally).

My monitor software (monit`s last version) didnt send me any alerts. Looks like the system crashed, but why I was able to ping it then??!

The logs is empty. No errors at all.

What that could be? Where I can check that? What should I do?
Did anybody got the same problem? Is that hardware problem?

My system is: dual p3, 1 gb ddr ram, ide 8mb cache drive. I have not heavy usage for such configuration: ~40k hosts/day.

Please help!
Thanks.

Sounds like bad ram, or some sort of kernel-level bug (i'll assume you use linux, as i've seen this happen with linux).

I'd check the ram and if you are running an older version of linux, upgrade to 2.4.20 (which was just released and has better SMP support, so you might want to do that anyways).

yes, im on linux 7.3
what pogram should I use for RAM check?
do you referring to glibs when you`re saying "2.4.20"? how can I check for its version on my box, instead of asking NOC?

thanks!

Sometimes, when a linux server runs out of swap space things like this happens.

Check for "do_swap" or any message related to swap.
cd /var/log
grep swap *|less

Solution: Usually to increase your RAM or to detect which processes uses more ram and try to remove them or move them to a second server.

Hope this helps
ernesto

I have 1GB DDR ram with this box, usualy 600-800Mb is free. Although its still use ~7-9Mb of swap space :confused:

I have experienced that problem with portsentry before.. somehow IPtables blocked off services on one server a few times, and I had to get the dc to reboot it or flush iptables.. disabled portsentry and it never happened again.. I could see portsentry was config'd correctly but still ended up dropping the machines IP's into iptables rules. Maybe have a look at something along those lines. :D

Exactly the same thing has happened to me.

Server was down (~6 hours) untill I got in contact with support, then they said: it's up again. Ok nice, but they didn't care to say WHY it crashed. If it happens again I am going to demand an explanation - I mean it is a managed dedicated so they're suppose to fix faulty memory (which I think is certainly one possible cause), cpus etc.

Server load is very low on my server (<2000 hits/day) so that's not the cause.

90 % kernel panic due to cpu

i was having the same problem

see you message log /var/log

and read the last lines after the system crach

you will see if you have a kernel panic or not