DL380 G3 Red Hat ES servers freeze

[Chris Coston]

We are running approx 20 RH Servers on RH 2.1 with 4GB's of RAM. Approx 8 of the 20 servers freeze every 3-15 days depending on server. Some are mail servers, application servers and My SQL servers. All are running the lastest HP drivers. In talking with HP, they are recommending that we remove SIM. There may be a bug in the most current agent. Is anyone else seeing a bug with SIM on DL380's with current versions of SIM?

Thanks,

[Steven E. Protter]

We have those servers and have not received any notices on that.

Kind of curious. Is the notice just for this version of Red Hat or all?

We're running RH ES 3.0 release 1.

SEP

[Roberto Polli · Accepted solution]

Hi,
got same problems using RH8 until upgraded to kernel 2.4.18-27.8.0smp. It was due to a bad NIC driver (tg3) which I changed with the bcm5700 taken from the HP site.

I use HP asm ver 6.4 because I experienced some problems after upgrading to v7.0

Which kernel are you using?

Peace,
R.

[Don_89]

I've had the same problems with Roberto with HPASM 7.0. We switched to ver 6.40 for the agents..

[Olivier Drouin]

From my experience with those machines:

If you use bcm5700 drivers and HP 6.40 agents and still experience the problem call HP to change the board under warranty...

Hardware problems I had with DLxxx machines:

dl360 g1: power supply fan (you will have a msg on console). I had a whole batch of these faulty fans.

bl20p: rIloe freezes sometimes (change the board). couple of blades.

dl380g2: changed some boards. Same problems as what you have...

...and maybe some others I dont remember...


If you dont have redundant setups all these annoying hardware problems can be a PITA.

Hey, its the joy of running x86!

[Chris Coston]

Reply to Steve.

We are actually working with a HP tech. They did not cite a notice or anything. Just that the tech had one other case with system hangs, removed SIM and remedied the problem. We have removed SIM and are still in the waiting peroid to see if any more system freezes happen. Incedently when the system does freeze, the sysrequest kernel dump commands did not work. One other interesting thing, the system freezes all happened in the evening around 9:pm until 6:am. We had approx 20 hangs over the last 3 months. We were able to eliminate applications as approx 1/3 the time the boxes froze during idle processing times.

Thanks all for your input.