From: Ken_Sedlacek@KYRUS.COM
Date: Wed Dec 04 2002 - 09:22:22 EST
snip: from Simon Green:
2. Extra care is needed when applying SSA microcode updates, to either
disks
or adapters. (Not a big deal, and all of the information is in the
instructions on the SSA site.)
I am in the process of doing some SSA adapter microcode updates on shared
SSA adapters/SSA disks.
Each H70 (4.3.3.10) has 2 SSA adapters sharing four 7133-D40 drawers.
The instructions specifically say that you need to update ALL SSA adapters
at the same time!
One if these H70s is our on-line Oracle DB.
The other H70 is our test Oracle server.
I was going to update just our test H70 SSA Adapters, but the instructions
say both H70s SSA adapters need to be done at the same time!
Does anyone have experience in this area?
Should I do the SSA adapter microcode update on both servers? or test H70
first, then on-line second??
Ken Sedlacek
AIX/TSM/UNIX Administrator
ksedlacek@kyrus.com
IBM Certified Specialist: RS/6000 AIX v4.3 Support
IBM Certified Specialist: RS/6000 SP & PSSP 3
Tivoli Certified Consultant - Tivoli Storage Manager v4.1
"Green, Simon"
<SGreen@KRAFTE To: aix-l@Princeton.EDU
UROPE.COM> cc:
Sent by: IBM Subject: Re: Question about pSeries and shared access to disks
AIX Discussion
List
<aix-l@Princet
on.EDU>
12/04/2002
05:38 AM
Please respond
to IBM AIX
Discussion
List
No problem to have a 7133 drawer in a loop between two, (or more) servers,
provided you don't want concurrent access to any disk(s). But you can use
some disks on one server and some on another.
There are some things to bear in mind, though.
1. All disks are visible in each node, so some care is needed to make sure
that you don't inadvertently extend a VG using another's system's disks.
This won't happen if the disk is in use, (VG varied on), but might do if
it's off line.
2. Extra care is needed when applying SSA microcode updates, to either
disks
or adapters. (Not a big deal, and all of the information is in the
instructions on the SSA site.)
3. Care is needed when rebooting. Powering off a system breaks the loop at
its adapter. With just two nodes it shouldn't be a problem, but with more
nodes or multiple adapters it is possible to cable the loop in such a way
that some disks get isolated when one node is rebooted.
If you have automated monitoring, you're going to get SSA Link errors every
time a node is rebooted, which may be annoying.
Do not reboot both nodes simultaneously, as there is a chance of SSA
microcode errors if both servers try to configure at the same time.
Assuming the same type of server, just wait a couple of minutes between; if
they're different boot the fastest one first.
Simon Green
Philip Morris ITSC Europe
AIX-L Archive at http://marc.theaimsgroup.com/?l=aix-l&r=1&w=2
AIX FAQ at http://www.faqs.org/faqs/aix-faq/
N.B. Unsolicited email from vendors will seldom be appreciated.
> -----Original Message-----
> From: Wesley Joyce [mailto:wjoyce@UVI.EDU]
> Sent: 03 December 2002 21:33
> To: aix-l@Princeton.EDU
> Subject: Question about pSeries and shared access to disks
>
>
> Is it possible for two pSeries servers to have shared access
> to each others
> 7133 drawer and SSA disks in a non HACMP setup?
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 22:16:23 EDT