help my hp ml 370 g3 hangs with linux

[vinay_17]

I had compiled linux kernel (2.4.18-24.7.x) with latest cciss patch for my HP ml370 g3 server. But after a day or so the system goes blank, Its screen goes back and no keyboard or any other activity can bring it back. needs a hard reboot. var/log/messages have nothing.

lokks like it goes in power saving kind of mode and stays there. (I tried with advanced power mangement support off/on in kernelbut no use.)

can some one please tell me what is going on with my machine.

thx in advance,
regd,
vinay

[Vitaly Karasik_1]

try bcm5700 driver for your NIC instead of tg3.

Regards,
Vitaly

[vinay_17]

Is it some know problem with tg3 ?or tg3 with HP ml 370 ? because same configuration works very fine for dell poweredge server for me.

Any ways i will try download and put bcm5700 and see if it works.
will keep everyone updated.

thx,
vinay

[vinay_17]

One more thing, it works fine through out with default red hat 7.3 kernels that come with release. ( modules.conf has tg3 driver and it runs for weeks, its just that 2.4.18-24.7.x kernel which is compiled by me gets stuck after a day or two)

[Vitaly Karasik_1 · Accepted solution]

I had this problem [system freezes every 1-2 days] with ml370 G3; see

http://h18000.www1.hp.com/support/files/server/us/download/19688.html

http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=103267

http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=79997

Regards,
Vitaly

[vinay_17]

Vitaly,
I went through the links that you sent me.And going through them it seemed that people are using "noapic" parameter for kernel boot.As bcm5700 source which i could find didnt seem updated for rh 7.3, I tried "noapic" option.
For last two days my machines seem to be working fine :)
I will monitor them for a couple of days and let you know if it worked.

thx a lot again
you are really helpfull

regds,
vinay

[vinay_17]

Hi,
My hpml 370 box is running for almost 4 days now without any problem (2.4.18-24.7.x with "noapic" option.)

As I was not sure what effect "noapic" will have on performance and all...I downloaded 2.4.18-27.7.x kernel source patched it up with cciss driver and ran it on other HP ML370 g3 box...it seems to be working fine without "noapic" option.
if this goes on for a week i will consider it fixed (it seems it will go on).

Thx a lot again, i wasn't able to figure out whether its problem with some buggy driver or my patching procedure, if it wasn't for your help.

regds,
vinay