ProLiant Servers - Netservers

LH4r Multiple problems

 
SOLVED
Go to solution
Philip Brooks
Occasional Contributor

LH4r Multiple problems

I've recently aquired a LH4r Netserver, and I am having a few problems firing it up.

1. the display on the front of the server locks up every now and then, it just simply freezes while im viewing through the logs, only way i found to unfreeze it is to remove the power plugs for a few seconds.

2. the logs are smeared in memory errors, the message "single bit memory error in slot 1" appear on most of them, i havent read them all due to there being around 500 of them. i'm guessing there is a problem with the RAM installed in the server, but i cant see anything wrong with it, during boot up the BIOS reads the memory fine, which leads me to my next problem

3. the bootup freezes after checking extended RAM, the amount it detected is the same as is installed, however i cannot enter setup or do anything as the boot freezes as soon as the checking of the extended RAM finishes.

I've checked these forums a bit and havent found anything relevant, if someone can point me to a post which could help me or simply explain whats going on that would be great
3 REPLIES 3
Robert
Advisor
Solution

Re: LH4r Multiple problems

It's not unusual for a machine to report the correct Ram size during boot and the Ram be defective.

There's a couple of things you can do to eliminate a few things.

You didn't say if you checked the RAM for proper mounting, dust accumulation, mismatched RAM, recently replaced RAM, or what.

This first step would have to be done anyway. Power off the machine and unplug it. Hold the power button in for about 5 seconds. This will discharge the mother board. Do a really close, and I mean close check for dust accumulation on the motherboard, and remove any dust found.

Next, remove each RAM stick and lay them out in order, careful not to handle the connecting pins. If dust is seen on the RAM or around their sockets, remove it, then reinstall all the sticks in their original order.

(Next) Hoping you have a floppy drive in the server, via another machine which also has a floppy drive, go to this Microsoft URL and download their free Ram test utility. http://oca.microsoft.com/en/windiag.asp

Once downloaded, insert an empty formatted floppy and execute the mtinst.exe from Microsoft. This little 640Kb execution creates a self booting diagnostic floppy.

In your post you said that the logs reported "single bit memory error in slot 1". This makes me believe that you are using at least two sticks of Ram.

Without knowing how many RAM sticks you're using, I'll have to generalize the possibilities for two sticks, and one processor.

Insert the diagnostic floppy you just created into the server and fire it up. Hopefully BIOS has been told to scan for a floppy during boot and if so, the diagnostic will immediately start running a 6 step per pass test. Each step will read "active" and the progress meters will be running. If a step passes it'll tell you that and proceed to the next test. When I test using this diagnostic, and it reports all 6 tests passed, I'll let it loop the tests for a least a dozen passes, hoping for the same results. This heats up the Ram and could expose heat related errors. If all tests pass, press "Control" + "X" to exit the diagnostic, which will immediately reboot the server.

In some cases, the diagnostic program will lock up as soon as it starts to run, in which case, one stick is definately bad, or there's a mismatched Ram stack, or improperly socketed Ram, etc.

If the diagnostic locks up or can't run at all, shut down, discharge the motherboard again, remove slot1's ram and re-run the diagnostic with the single ram in slot0. If it reports good, slot1's ram is bad. You can confirm this by powering down again, discharge again, and put the suspected bad ram into slot0 as a single, and run the diagnostic again. This would confirm that the Ram stick formerly in slot1, tested bad again by itself in slot0.

Then it's a matter of replacing the Ram. It would be good to know that the Ram you're testing is the original Ram for the server (check the specs). If any ram stick has been replaced in the past, make sure the working Ram is the correct Ram for the system, and replace the bad one with what's recommended for your server. If the bad Ram is not what was spec'd for the server, enough said. Good Lock. I'll be watching your post with curiosity and additional help if needed.
Philip Brooks
Occasional Contributor

Re: LH4r Multiple problems

Thanks a bunch, turns out a few of the RAM sticks were a different size.

that solved pretty much everything I addressed in my first post, even the front display hasnt locked up.

ill post here again if anything breaks on me.

fyi: it has 4x 500mhz and had 16 sticks of RAM

thanks again.
Robert
Advisor

Re: LH4r Multiple problems

Good Deal. Use The Ram test anyway and let it check for faults. Not a bad thing to know that your Ram Tests Good. Have Fun.