Comments for Vallard's Tech Notes

Comment on x3650 M2 boot issue with QLogic HBAs by Alexandre

Alexandre — Wed, 25 Aug 2010 21:09:00 +0000

http://www-947.ibm.com/support/entry/portal/docdisplay?brandind=5000008&lndocid=MIGR-5081613

Comment on Creating USB key to install VMware ESXi 4.1 by roadrunner

roadrunner — Sun, 22 Aug 2010 13:32:00 +0000

cool – works fine. Thank you for sharing the procedure.

Comment on Compiling NWChem by Anonymous

Anonymous — Thu, 19 Aug 2010 22:12:00 +0000

Test my comments.

Comment on Creating USB key to install VMware ESXi 4.1 by Procédure pour créer une clé USB avec VMware vSphere Hypervisor 4.1 (ESXi) | Blog-Note

Mon, 02 Aug 2010 08:35:47 +0000

[...] Pour les utilisateurs sous Linux, la procédure d’installation d’un vSphere Hypervisor 4.1 sur clé USB peut se lire ici. [...]

Comment on New pastures by Denis Fateyev

Denis Fateyev — Fri, 29 Jan 2010 19:22:28 +0000

Good luck, Vallard!
Previously, I read your articles on IBM website, they’re very interesting (and even helped me in something).

Comment on x3650 M2 boot issue with QLogic HBAs by YC

YC — Thu, 21 Jan 2010 10:00:41 +0000

Well IBM claims that “For non-UEFI aware OSes, this behavior is by-design”

Local operating system boot fails when external storage is attached – IBM System x3550 M2, x3650 M2
http://www-947.ibm.com/systems/support/supportsite.wss/docdisplay?lndocid=MIGR-5081613&brandind=5000008

Comment on updating iDataPlex IMM firmware by Vallard

Vallard — Mon, 04 Jan 2010 23:34:59 +0000

It does appear that after applying this particular firmware update the user ID and password will be set to the default. This happens in both instances.

Comment on dx360 M2 BIOS updates and manipulations by Vallard

Vallard — Tue, 15 Dec 2009 20:05:58 +0000

We have also been seeing strange problems that development has not given us any answer to. For example, our nodes have the same BIOS and IMM firmware level, yet give different results:

Some nodes are set to:
uEFI.OperatingMode=Custom Mode
while others are set to
uEFI.Operating Mode=Custom

asu showvalues is different too:
uEFI.Operating Mode=Minimum Acoustic/Power Mode==Performance Mode=Custom
vs
uEFI.OperatingMode==Acoustic Mode=Efficiency Mode=Performance Mode

Another example is uEFI.PackageCState
uEFI.PackageCState=C3 state vs uEFI.PackageCState=ACPI C3
Again, “showvalues” does match the settings
uEFI.PackageCState=C0 state=C1 state=C3 state==C7 state=No Limit
vs
uEFI.PackageCState=C0 state=C1 state=ACPI C2==C7 state=No Limit

We’ve been told to upgrade the firmware and determine whether or not this helps.

As far as documentation, I have not been privy to any, and sure wish I could.

My thoughts:
1. Reseting to factory defaults, yanking the planars, and then running ASU seems to yield better results.
2. I’m leary about running ASU on the node itself, butter to run ASU remotely using the –host option.

Comment on dx360 M2 BIOS updates and manipulations by John Walsh

John Walsh — Fri, 04 Dec 2009 01:11:14 +0000

Is there some documentation somewhere that describes the following settings:

set uEFI.TurboModeEnable Enable
set uEFI.AdjacentCacheLinePrefetch Disable
set uEFI.OperatingMode “Efficiency Mode”

There seems to be a variance in our iDataPlex cluster and some nodes are are set with TurboMode Enabled, others disabled. Most of them are also set to OperatingMode = “Custom Mode.”

The problem that I am seeing is that on a handful of the nodes we are seeing a lower memory performance then others. Streams reports back on average about 28 GB/s but on a handful of nodes it is around 20% lower or around 24 GB/s. Any suggestions or thoughts?

Comment on IPMI Settings by Vallard

Vallard — Tue, 01 Dec 2009 00:31:52 +0000

IBM iDataPlex dx360 M2