[KLUG Members] Attempting to fix a server

[Adam Bultman]

Phillip Hofmeister wrote:

>On Thu, 26 Aug 2004 at 09:22:12AM -0400, Adam Tauno WIlliams wrote:
>  
>
>>>Last week, I had a server tip over on me.   I believe that the SCSI card 
>>>had pretty much up and died, and as a result we had some filesystem 
>>>corruption, and the inability to boot.
>>>At the time, the system would 'start' to boot, and then reboot when it 
>>>started getting too far, and start over again.
>>>      
>>>
>>Have you tried a memory test?  Reseated any DIMMs/SIMMs.  This sound
>>allot like a potential memory problem.
>>    
>>
>
>I had a computer starting to act up a year ago.  I ran memtest86 (runs
>out of lilo or your boot loader).  It quickly identified a problem in an
>address area in my second DIMM.  Removed the DIMM, everything worked
>fine again.
>
>  
>
The system, before it REALLY tanked, would crash somewhat often, and 
mostly with SCSI errorsduring large transfers of data.  It was an NFS 
server for all sorts of data, and would be fine, until you backed it up 
too fast or tried to copy too much data at once.  Then it would cry like 
a little girl and crash.  Unfortunately, it's a 1u server, with no CDROM 
and no floppy, so I don't know how I'd go about running memtest86 on it 
(unless I netbooted it, ugh).

I'll be working on it again tonight (gonna replace the password/shadow 
files, probably). If it runs, I'll see what I can do aobut swapping the 
RAM and testing the RAM in another machine.

Adam