Tales from the Datacenter » IBM http://www.pburch.com/blog Tales from the Datacenter Thu, 10 Jun 2010 17:51:08 +0000 en hourly 1 http://wordpress.org/?v=3.0 Best Practice: Cabling an EXP810 to a DS4700 http://www.pburch.com/blog/2009/08/04/best-practice-cabling-an-exp810-to-a-ds4700/ http://www.pburch.com/blog/2009/08/04/best-practice-cabling-an-exp810-to-a-ds4700/#comments Tue, 04 Aug 2009 13:46:00 +0000 Patrick http://www.pburch.com/blog/2009/08/04/best-practice-cabling-an-exp810-to-a-ds4700/ We’ve been expanding our storage capacity here at work in the recent weeks.  As such, cabling all of these IBM DS4700s and their associated EXP810s (full of glorious fiber channel and SATA drives) became a task.

The question has always been how best to cable the fiber that keeps all this stuff working properly.  You see, each DS4700 (the controlling enclosure) has two fiber controllers, each with two HBAs for connecting the EXP810 expansion enclosures.  Well, we came across an IBM document that details exactly how they recommend cabling – and here it is.  And here is the recommendation for a DS4700 with its maximum number (six) of expansion EXP810s (click for a larger view):

ds4700exp810cable

]]> http://www.pburch.com/blog/2009/08/04/best-practice-cabling-an-exp810-to-a-ds4700/feed/ 0 Microcode Levels http://www.pburch.com/blog/2009/02/10/microcode-levels/ http://www.pburch.com/blog/2009/02/10/microcode-levels/#comments Wed, 11 Feb 2009 02:35:37 +0000 Patrick http://www.pburch.com/blog/?p=53 I was configuring our twelfth virtual host this afternoon and ran across a weird issue.  You see, we buy IBM Blade servers that go through integration, which means someone else configures the hardware for us.  Unfortunately, I don’t think they turn the hardware on and check things.  Case in point: we had Blades with processors running two different microcode levels (think firmware for a processor).

Little did I know anything was wrong when I got a big fat blue STOP error.  MULTIPROCESSOR_CONFIGURATION_NOT_SUPPORTED was staring me in the face.

Remember, check your microcode levels if you see STOP 0x0000003E.  It could save your life.

]]> http://www.pburch.com/blog/2009/02/10/microcode-levels/feed/ 0 Management Ports Gone Wild http://www.pburch.com/blog/2008/06/27/management-ports-gone-wild/ http://www.pburch.com/blog/2008/06/27/management-ports-gone-wild/#comments Sat, 28 Jun 2008 01:12:13 +0000 Patrick http://pburch.com/blog/?p=9 Today, I learned a very important lesson.  That lesson was that management ports on IBM xSeries server do not provide network connectivity (at least not by default anyway).  That’s right.  Huge news flash, right?

Actually, the lesson was to have the technician on the other end of the phone make sure they aren’t plugging into the management port.  Now, don’t take that the wrong way.  He didn’t know any better than I and merely plugged the patch cable into the only RJ-45 port he could see.

So, as a reminder to myself to check the management port, I blog.  Let that be a lesson to you.  Then again, maybe we should ask ourselves what prompted the network connection to drop in the first place?  I’ve already opened a ticket with IBM, because we thought the network port was dead.  To that end, I’ll run DSA on Monday and see if IBM’s “Remote Technical Support in Atlanta, Georgia” can tell me anything about it.

]]> http://www.pburch.com/blog/2008/06/27/management-ports-gone-wild/feed/ 0