Server sometimes (1 or 2 hrs) down :/

[edge]

Originally Posted by falko

At least it's possible.
Do you get the same problems if you leave the firewall off? That would mean it isn't caused by the firewall.

Hmm I do not really want to leave the firewall off, and I can not really test it as this error sometimes does not happen for days/weeks!

I've been google'ing about the 'sk98lin', but did not find any good info.. What does it do?

[falko]

Originally Posted by edge

I've been google'ing about the 'sk98lin', but did not find any good info.. What does it do?

It's a network card driver:
http://kerneltrap.org/node/3419
http://gentoo-wiki.com/HARDWARE_sk98lin
http://www.die.net/doc/linux/man/man4/sk98lin.4.html

[dusti]

I have these strange isues with a dell GB nic also...
The NIC is working fine most of the time and then suddenly all goes wrong.

I did some pings and arping while capturing output with tcpdump -v -v
that was interresting:
If I pinged to annother server or the gateway It would do it good for about 5 times then it goes all wrong for about 50 pings, then it pings again wright for about 5 times and on and on...
seemed that for some obsure season the the NIC forgot what the mac was of the machine it was talking too, so it started broadcasting, gets his answer, pings 5 times and then forgets it again?
When I arping'd it just kept going...
To me it seems like a driver issue, specialy since the same configuration works flawless with a 3 com GB pci NIC. Don't know if it's related to the PCI-X or with the NIC itself...



divert: allocating divert_blk for eth0
eth0: Tigon3 [partno(BCM95721) rev 4101 PHY(5750)] (PCI Express) 10/100/1000BaseT Ethernet 00:15:c5:5f:03:96
eth0: RXcsums[1] LinkChgREG[1] MIirq[1] ASF[1] Split[0] WireSpeed[1] TSOcap[1]
eth0: dma_rwctrl[76180000]

divert: allocating divert_blk for eth1
eth1: 3Com Gigabit NIC (3C2000)

Uname -a
Linux elmo.domain.xx 2.6.9-34.ELsmp #1 SMP Wed Mar 8 00:27:03 CST 2006 i686 i686 i386 GNU/Linux

I have no clue what or how I could test further to find out what the problem really is (and how I could solve it...)

[edge]

dusti,

Are you using any virtual IP's / NIC's? If so.. remove all of them (only use the main NIC with one IP), and see if you still have the problem. If not, add one virtual IP / NIC, and see if all is still okay. etc ect.. I think that in my case one of the virtual IP's was causing the problem.

An other thing I did (as backup) was that I made a script pinging the gateway ever 5 minutes.. When it did not get a reply, it did a network restart.. After the network restart the network was working fine again.

btw. I'm now not using Fedora anymore (it's Debian now), and all is working fine with my Dell PowerEdge and it's 16 IP's

[dusti]

I am not using virtual IP's, at the moment I only need 1,
Later this week I'll try the driver rpm I got with the machine, see what that gives. I know debian is sometimes better, somethimes fedora is better,
anyway, I just have to run fedora/redhat-based here.
I also could try to add a virtual nic and see if that one is doing a better job

[edge]

dusti,

please do not get me wrong with the Fedora v/s Debian. I have an other server running Fedora, and both are great OS'es!

The reason I installed Debian on that server was only to try an other OS!