| Author |
Message
|
| csmith28 |
 Posted: Wed Jul 13, 2005 4:32 pm Post subject: This is a Classic...... |
 |
|
Grand Master
Joined: 15 Jul 2003
Posts: 1196
Location: Arizona
|
Today, one of our new, superior technology Hitachi Blade Servers running some Flavor of Linux went into YO-YO Mode rebooting itself at random intervals.
Eventually the reason was discovered. Apparently between the vibrations of the CPU cooling fan and the thrust of the fan, one of the CPU's was literally lifted out of it's socket. 
So when the server shutdown the CPU Fan would stop and the CPU would then rest back down into it's socket just enough to allow the server to come back up until the force of the fan yet again lifted it far enough out of the socket for it to lose connectivity to its leads causing the server to panic and shutdown again. 
_________________
Yes, I am an agent of Satan but my duties are largely ceremonial.
Last edited by csmith28 on Wed Jul 13, 2005 4:39 pm; edited 1 time in total |
|
| Back to top |
|
 |
| PeterPotkay |
| Posted: Wed Jul 13, 2005 4:37 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
Last month one of my WB-IMB broker servers kept rebooting randomly. Not until they replaced the CPUs did it become stable. Hmmmmm, I wonder.....
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| csmith28 |
| Posted: Wed Jul 13, 2005 4:42 pm Post subject: |
|
|
Grand Master
Joined: 15 Jul 2003
Posts: 1196
Location: Arizona
|
Was it a Blade Server or one of the more traditional Enterprise Server types.
Generally speaking the Enterprise Server CPU fans are mounted so they blow on the top of the CPU but in most blades due to space restrictions the fans blow on the CPU's from the side.
_________________
Yes, I am an agent of Satan but my duties are largely ceremonial. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Wed Jul 13, 2005 4:44 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
It was not a blade. Still, I will mention your story to my server guys...
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| malammik |
| Posted: Thu Jul 14, 2005 4:55 am Post subject: |
|
|
Partisan
Joined: 27 Jan 2005
Posts: 397
Location: Philadelphia, PA
|
I think its a good pracitice to burn in new servers before loading anything important on them. My friend works for a hosting company. After the server is installed they keep re-compiling the kernel on it, generate large rsa keys and run IO benchmarks for 3 days or so. Sometimes you get a loose processor, sometime you get bad memory, sometimes a damaged disk. If think if someone actually did a statistical analisys he would find out that servers are most likely to fail within first few days in the server room AND few years later when the parts actually begin to wear out.
_________________
Mikhail Malamud
http://www.netflexity.com
http://groups.google.com/group/qflex |
|
| Back to top |
|
|
| scottm |
| Posted: Thu Jul 14, 2005 9:02 am Post subject: |
|
|
Apprentice
Joined: 13 May 2004
Posts: 44
Location: SE Tennessee
|
And did they say it was an MQ problem???  |
|
| Back to top |
|
|
| csmith28 |
| Posted: Thu Jul 14, 2005 1:34 pm Post subject: |
|
|
Grand Master
Joined: 15 Jul 2003
Posts: 1196
Location: Arizona
|
| scottm wrote: |
| And did they say it was an MQ problem??? |
Well I'm not sure about Peter's case but in my case no.
The server it happened on where I work was a WebLogic Server. I don't have any MQManagers hosted on Blade/Linux Servers. 
_________________
Yes, I am an agent of Satan but my duties are largely ceremonial. |
|
| Back to top |
|
|
| sebastianhirt |
| Posted: Thu Jul 14, 2005 2:12 pm Post subject: |
|
|
Yatiri
Joined: 07 Jun 2004
Posts: 620
Location: Germany
|
Unbelivable how much server they get in such a little blade, though.
But they say the first super computers had about the performance of a modern scientific calculator. I am really looking forward to see where we are with technology in 10 Years time. |
|
| Back to top |
|
|
|
|
