ProLiant 5500 locks up or ASR's every day or so

[Charles Cosby_3]

Ok I have asked this question about this machine before and still have yet to cure the problem. I have a ProLiant 5500 with dual PIII 550 Xeon 1 meg processors. No matter what OS I have tried (W2K server, advanced and windows 2003 standard) it runs for a day or so (sometimes as long as 48 hours) then locks up or does an ASR. I have tried wiping the system many many times, tried everything I can think of, updated the bios on everything and nothing seems to help. Last week I wiped the entire system and did a fresh install of Windows 2003 Standard Server (microsoft loves all my phone calls to reactivate). I stripped down the hardware to bare minimum (4 x 64 meg ram was the smallest I have for it). I literally did nothing but install the OS.

The only problem I had was that it did not recognize my NIC (dual head netintelligent). I installed those drivers and it worked fine. Ran for about 22 hours and locked up. I swapped out the ram for others, and again about a day and lockup. I did that through all 4 sets of ram and have confirmed it is not a ram problem. I ran HP's diagnostics and ran everything, it found absolutely nothing wrong. Left it for a day and sure enough, lockup.

Then I was noticing that every time it is locked up, the one array has all the activity lights lit up on the drives attached to it every time. I have a SMART 3200 RAID controller. I have it set up with 2 - 9.1 Gig mirrored for boot and the other 6 - 9.1 gig as RAID 5. It is the RAID 5 that lights up when it locks up. I have the 2 sides of the cage attached to the 2 internal channels on the card. Yes it is a duplex cage.

Is there a problem having this array over both channels? Also would it be a RAID controller fault if the lights stay on like that? As you can imagine, I am quite frustrated with this machine. This has been going on for over a year and I am beginning to lose patience.

The only other thing I thought of was swapping out CPU's or taking one out. Had to do that once with a Proliant 4000. Problem is I dont have a blank or spare CPU to swap.

Thanks for any advice or help you can offer.
Charles

[Leon Rosier · Accepted solution]

The lights on your disks is normal behavior.
When the array controller is idle, the lights on one array all off and on the other one they are all on. I would expect the standard peripheral board first (where the video is connected through). The system board and processors showed to be very reliable. Also try to run hpdiag without your nic controller or replace it by another one.

Leon

[Charles Cosby_3]

Leon,

Thank you, thank you, thank you.

I fought with this beast for almost a year and it never dawned on me to replace the NIC. Swapped it out and now its been running great!!!! Just wish I could reward more than 10 points!!

Thanks again
Charles