IBM server 3950 M2 server hang

Dear All,

  I have two IBM server 3950 M2 connected to each other via IBM Cascade cables ( Expansion ), sometimes the server stops responding, means I can not ping it , when I go physically I see the windows login screen, when I login its just showing welcome message, I had to restart the server. this issue happens three times, I checked the event logs but can not find any errors related to it, can someone advise how to solve thus issue. the server is 8 way server with windows 2008 R2 enterprise with 150 Gb RAM.


Thanks
LVL 1
ITMaster1979Asked:
Who is Participating?
 
DavidConnect With a Mentor PresidentCommented:
Well, with intermittent problems on an expensive host then you pretty much only have a few options.

1. run IBMs full diagnostics and hope you get lucky and they find something.
2. Swap out bits of memory and other components over the next day, week, or month and hope you get lucky and find something.
3. Buy motherboard and PSU test hardware designed specifically for the purpose of testing flaky hardware (or take it to somebody that has the gear).

There are so many things that can explain this, and with an 8way W2K8 server you probably aren't in a position to do brute force diagnostics.

(Personally, I am a big fan of #2. Decent test hardware can be bought used on ebay for under $500).
0
 
DavidPresidentCommented:
If you can afford a lot of down time, boot the system to another o/s, like LINUX installed on a USB stick, and set it up to play videos on youtube or some other site. If it still crashes and locks up, then you at least know it is unlikely to be a problem specific to windows.

Booting LINUX is safe. It won't mount any existing windows partitions in read-write mode, so you can even do read tests safely by copying raw data from those disks into the bit bucket

i.e, dd if=/dev/sdb  of=/dev/null bs=64k &
(This copies from /dev/sdb into nothing 64KB at a time, and runs in background.  If you boot to a USB stick then the USB will be /dev/sda.
0
 
ITMaster1979Author Commented:
I noticed it happens every week at the same time . Any other advise
0
 
DavidPresidentCommented:
At the SAME time?  Then look at the O/S and see what automated scheduled tasks run every 2 weeks prior to the crash.

Manipulate the scheduling of them, so the first one runs at lunch time, after giving everybody fair warning that the system may lock up.  If system doesn't lock up, keep trying.

Not everything may be listed, example Adobe checks for updates every X days at Y time, and this is configurable via a separate program.  But still, you get the idea.  You are lucky that you know WHEN you will have a crash.  At very least you can turn on full system logging before the time system is to crash and maybe one of them will tell you what happened.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.