Server Problems
 

I have an IBM Netfinity 5100 Server set up in my house. I know it is serious overkil for just my wife and I, but what the he!!, it was free! :)

Anyway, all was working fine up until a few weeks ago, when it basically just shut itself down and would not boot up anymore. The fault indicator lights inside told me the 733 MHz CPU in it died. It's a Pentium 3 board, and I happened to have a spare, but slower processor (533 MHz), so I put it in, booted up and was able to work with it. Unfortunately, the DVD burner I have installed in it requires a faster CPU, so I was unable to perform my backups.

I ordered a couple of 733 MHz processors for it and when they arrived, I started the server and configured it for dual processors. Shut it down, installed both processors and basically have not been able to boot it since.

Initially, the POST would tell me that CPU 2 failed POST. Swapped the two processors and continued to get the same error. Further troubleshooting revealed that I needed to install a voltage regulator to run dual processors. Ordered the VR, installed it and tried to boot. I no longer received the CPU failed POST error. In fact, I got nuthin!!! Do display showing up on the monitor.

I've removed the VR, installed one processor at a time and still cannot get anything to show up on the monitor. Removed and replaced the RAM, alternated CPUs, confirmed the monitor works by connecting it to another pc, and still cannot get a display.

Help!?

Randy

Are the CPUs the same stepping?

Try to come up on one at a time, and see if the BIOS will show you the CPU ID. If they are different, they will not work together.

Yep, processors are exactly identical.

I've tried them together, swapped them in the slots, installed one at a time...nuthin!

Randy

You killed it.


I'm not sure if IBM does like we do (I work for Dell doing server support).... you could try to call their tech support and give them the rundown and see if what you did toasted the system board or some other component.

Are you getting a beep sequence when you power on?

WOW!!!! I believe I've just witnessed a resurrection!!!

Playing around with it again, while typing my posts here. Removed everything again and started re-installing components one at a time and magically, it's working again. All components are back in and it seems to be working like a champ!

edit: Although it is back to running off of one processor, the server is still set up as a multi-processor system. Gonna have to fix that, I suppose. I'll try the removed processor in another system to see if it has been fried or not.

Randy

Matt,

Appreciate the help, buddy!

Brandon,

I am now! :)

Randy

IBM's weird that way.

We have to play with about 20% of the boxes that come into my DC in Dallas the same way - unplug power up, plug, power up, move part, power up, etc...until something works.

Reseating components fixes a great amount of things! I have my customers do that many times, and it resolves issues quite often...

Well, looks like I won't be testing this processor in another box. The suspect processor is a Slot 1 and my other P3 machines are Socket 370's. Maybe I can find an old pc at work to mess with.

If not, I'll give my server a few days and then will try adding the second processor again.

And yes, reseating components can be your friend!

Randy

I have to say that I have a few IBM servers and I HATE them with every fibber of my body.
They have been purchased and installed at different times (so it's not just a bad batch), they are all different models of the xServer series and they ALL have had hardware issues from purchase.
The first two took an IBM techo about a month to get running stable and the third took 6 months.
IBM tech support is a nightmare to get to the correct person (someone who understands what you are saying) and they don't trust the end user to have a valid opinion.
me: I think it's the raid card that's crashing.
them: hardware diagnostics says there are no errors so it must be the drivers for your OS (they told us to use redhat 7.3)
((side rant - at the time if you looked at their site they have a page that says what OS is supported and another page that says what OS is certified. redhat 7.3 was certified but not supported and redhat 8 was supported but not certified!))
me: (installed redhat 7.3 with about 10 different versions of the driver (that they kept sending me)) still doesn't work.
them: must be the OS.. use redhat 8
me: (installed redhat 8 with all the different versions of the drivers) still doesn't work
them: redhat enterprise is the only one that works, you must buy that.
me: (checks the price and it's the same as windows 2003) ***** it, I'll convert to windows and then they better not say it's the driver.
(installed windows 2003 and it still crashes) HEY, LIKE I SAID BEFORE, IT'S THE BLOODY RAID CARD.
them: they finally replace that card (and some others) and wow... it works.

Never again.

p.s. Sorry about the rant.