Intel Communities: Message List - IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

mgregg@redhat.com — Fri, 26 Oct 2012 18:58:48 GMT

I received 20 x3650 M4's last week.

I think that my feedback would be about the time it takes for the system to post. It is incredibly long.

We have a test system that these machines are supposed to be for that reboots and re-provisions these machines several times a day.

These systems will typically take greater than 5 minutes to finish post. This is arduous to wait for. They even take longer that a minute or more to even power on.

Is there a way to disable most of what they are doing to get to a sub minute post time?

I have called IBM support about this already, and they seem to suggest there is nothing that I can do about this.

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

d3xx@hotmail.com — Wed, 01 Aug 2012 07:51:38 GMT

I'm installing Centos on a new IBM x3250 M4. The machine itself is nice. However ive never seen a server take sooo long to boot! Its a good 6 minutes before it even starts to load from cd or installed os. Surely IBM management modules shouldnt take that long?

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

jon.strabala@quantumsi.com — Mon, 06 Feb 2012 22:10:47 GMT

I am using the following on a new IBM x3550 M3 (12GB ram, 2.266GHz Xenon processor)

Firmware Type Version String Release Date
IMM YUOOC7E 09/30/2011
UEFI D6E154A 09/23/2011
DSA DSYT89P 10/28/2011

Firmware Type	Version String	Release Date
IMM	YUOOC7E	09/30/2011
UEFI	D6E154A	09/23/2011
DSA	DSYT89P	10/28/2011

I must say I am sorely dissapointed with the "speed" of USB booting in the legacy BIOS mode in the IBM UEFI implementation.

Basically a "similar CPU" in a Sun X4275 will boot a 275 MB usb key image in just 32 seconds, while the IBM x3550 M3 takes over 363 seconds for the same image. Measuring the IBM form the time it starts a legacy USB key boot until I get an OS prompt is ridiculously long.

BEG: 1:27:05 pm (start SmartOS USB 2.0 USB key)

END: 1:33:38 pm (done into running Solaris 11)

---

TOOK: 6:33 (six minutes and 33 seconds - pretty slow - only 0.75MB/sec.)

It is almost as if the UEFI implementation uses a tiny block size like 512 byte read, rather than a larger buffer during reads. Once I am in the OS I can benchmark the performance of the USB key I booted off, IMHO if the IBM UEFI code read a 8192 or better yet a 32768 block size booting would be super fast.

So in a Solaris 11 operating systems we see the following performance characteristics for my USB key, ranging form 512 byte to 131072 bytes. Looks like either 8192 (12.3 MB/sec in a booted OS) or better yet a 32768 (20.2 MB/sec in a booted OS) would be a nice read size. It also looks like a 512 block size (0.64 MB/sec in a booted OS) matches the results I seem to experience in my lengthy boots.

time dd if=/dev/dsk/c1t0d0p0 of=/dev/null bs=512 count=524288

524288+0 records in

524288+0 records out

real 31m19.499s

=> 00.64MB/sec. on Solaris 11 (this is the speed of the IBM bios boot speed)

time dd if=/dev/dsk/c1t0d0p0 of=/dev/null bs=1024 count=262144

262144+0 records in

262144+0 records out

real 1m39.989s

=> 02.56MB/sec.Solaris 11

time dd if=/dev/dsk/c1t0d0p0 of=/dev/null bs=2048 count=131072

131072+0 records in

131072+0 records out

real 0m50.215s

=> 05.09MB/sec. Solaris 11

time dd if=/dev/dsk/c1t0d0p0 of=/dev/null bs=4096 count=65536

65536+0 records in

65536+0 records out

real 0m33.056s

=> 07.74MB/sec. Solaris 11

time dd if=/dev/dsk/c1t0d0p0 of=/dev/null bs=8192 count=32768

32768+0 records in

32768+0 records out

real 0m20.757s

=> 12.33MB/sec. Solaris 11

time dd if=/dev/dsk/c1t0d0p0 of=/dev/null bs=32768 count=8192

8192+0 records in

8192+0 records out

real 0m12.785s

=> 20.02MB/sec. on smartos (as expeected and seen on a Win7 box)

time dd if=/dev/dsk/c1t0d0p0 of=/dev/null bs=131072 count=2048

2048+0 records in

2048+0 records out

real 0m11.532s

=> 22.19MB/sec. Solaris 11

Making a UEFI compliant USB key might be possible but seems like a lot of effort, and of course the "same slow" USB key read issue might also exist in the UEFI implementation. Others complain about slow UEFI boot from USB keys refer to google searches like:

ESXi 4 hypervisor slow usb boot +IBM +"system x"

+IBM +"system x" slow usb boot

The only reference to a work around is a cryptic document section "XSW02525-USEN-00", or Introducing UEFI-Compliant Firmware on IBM System x and BladeCenter Servers, which doesn't really say anything useful.

Jon Strabala

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

matthew@rampant.com.au — Tue, 14 Dec 2010 10:49:37 GMT

I would like to agree with Unbiased on the stupidity. Of this supposedly new age system.

I have had one dumped on me to remove VMWare and put Server 2008 on it.

The thing will not boot off the DVD Unless it is a VMWare DVD.

After hours of thralling through IBM's useless junk I hit Google, After at east an hour of reading others with the same Issue I find this http://www-947.ibm.com/support/entry/portal/docdisplay?lndocid=MIGR-5085881&brandind=5000020 Which tells me the install two MS apps before I install mt new OS. Now being an MS tech I do know how to do this. But where on this page is IBM's account details for issueing the invoice for the time I had to waste due to their exceptionally intersesting systems and installations techniques. Quite frankly any Solution that any company pinned to the WWW that required the Client to hire an MS certified tech to spend hours and hours fixing their own buggy junk should be disbarred from ever selling sdaid junk again.

I realise this post will go wholly unnoticed but that will not stop me bagging the manufactures of this buggy junk till they stop or go bankrupt. I think the funny thing this time is the owner of said system happens to be a very large IBM dealer. When I explain what the issue is I am sure they will annoy there upstream people till they get some thing for free. So this one will cost you IBM, It will cost me too but at least it will cost you.

Now off to see if I can find a way of fixing IBM's stuff up without having to dismantle rack and return to workshop or spend hours in car back and forth trying to get the boot version right.

Hope others read this and realise they to have been duped.

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

webadmin@intel.com — Wed, 23 Jun 2010 00:52:15 GMT

Critical updates released 12 months after market release are UNACCEPTABLE

And bios/raid interface one of the most user unfriendly I have seen in 20 years.

Attempting to download siad updates from 3 different sites also timed out. You have inadequate bandwidth for the job.

And this is my second attempt to write this as i didn't get the verify code right and you guys couldn't even write a session variable to hold my comments.

Overall an extemely disappointing result

Critical updates released 12 months after market release are UNACCEPTABLE

And bios/raid interface one of the most user unfriendly I have seen in 20 years.

Attempting to download siad updates from 3 different sites also timed out. You have inadequate bandwidth for the job.

And this is my second attempt to write this, as i didn't get the verify code right and you guys couldn't even write a session variable to hold my comments.

Overall an extemely disappointing result

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

webadmin@intel.com — Fri, 16 Apr 2010 00:57:20 GMT

Hi Mike;

I have tried to install ESX 4.0 on IBM x3550 M2 server. As ESX 4.0 is non-uEFI aware OS, I have
setup "Legacy Only" flag as per the white paper.For some reasons, the system failed to boot
from CD/DVD Rom.

The boot manager is setup as follows:

Legacy Only
CD/DVD Rom
Floppy
Hard Disk0

I am sure the DVD media is good because I can be able to install on other system. Also
the system can successfully boot from Windows 2008 (uEFI aware OS) DVD.

Please advise

Thanks
-Soe

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

webadmin@intel.com — Sat, 20 Feb 2010 19:28:22 GMT

We have just release a new white paper to help people tune their UEFI based system to speed up boot time. Please review this document:

http://www-947.ibm.com/systems/support/supportsite.wss/docdisplay?lndocid=MIGR-5083207&brandind=5000020

Regards,

Michael Brinkman

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

webadmin@intel.com — Wed, 21 Oct 2009 15:40:13 GMT

Just got done getting the HS22 to Boot From SAN (SVC). We recently received a HS22 blade (7870-AC1) and tried to configure it to Boot From SAN. After several attempts of trying to configure the Qlogic card (QMI2572) a support call was placed to IBM since the Qlogic card was unable to save any settings. Another Qlogic card was also ordered to rule out a defective card and it had the same issue. I wouldn't even recommend trying to Boot From SAN without these firmware levels.

Firmware levels should be at:
BIOS 1.04, Build P9E130AUS
Diagnostics 1.13, Build P9YT40A
Blade Sys Mgmt Processor 1.05
Qlogic QMI2572, BIOS Revision 2.08
IMM (Integrated Management Module) 1.05, Build YU0032F

Overall we were able to configure the HS22 to Boot From SAN but were only able to do so by loading an OS (internal drives) and applying the updates since Update Express didn't have the latest firmware for the Qlogic card. The time to deploy a blade has increased. The HS21 took less than 30 minutes to have up and runnning when booting from the SAN compared to the 2 1/2 hrs to deploy the HS22. Please share this email as needed.IBM x3550 M2, x3650 M2, HS22, and dx360

Message was edited by: William Lea, adjusted to smaller font and removed Bold.

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

webadmin@intel.com — Mon, 19 Oct 2009 23:27:38 GMT

Hi,

working on a couple of x3550 M2's in a remote co-location, dont have kvm hardware, or ibm director in this case - so have been relying on the remote admin features of imm. would have loved to use the remote console but the requirements needed to get a java webstart app going are impossible for me, ie that the host im running the browser on can't get to the internet, and due to site rules there's no chance of fixing that - so the remote console feature is basically useless for now, could someone consider this scenario maybe come up with something that doesnt need to pull .jar files down from t'internet? ... anyway, we're using redhat, so we have the option of a serial console at least and I did figure out how com2 is accessible via the cli imm - 'console 1' command, so i prevailed.

Our host was shipped to us 3 weeks ago and had such old firmware on it that it really wasnt workable, perhaps you could push an alert up the supply chain to try and ensure units are not shipped out with known bad firmware - i spent the best part of a day with hung imm, that died simply because i rebooted the host OS, arranging remote staff to powercycle boxes, applying firmware and waiting 5-15 minutes each restart until i got them to a stable configuration - it's not the sort of experience that motivates customers to buy another 100 or so units.

anyway - now the servers are at these levels....

IMM	YUOO32F-2009/08/26	08/26/2009
UEFI	D6E128A-2009/08/20	08/20/2009
DSA	DSYT19A-2009/08/20	08/20/2009

able to boot/build/manage ok - but looking forward to your speedups in Q4 release.. - hopefully you can do something with the warm-reboot times - there seems to be about 5 minutes of dead time during a simple reboot that is not desirable. maybe if i was looking at the console i'd see what the holdup is...

one other thing i've noticed is that the 'onetime PXE Network Boot' option on the http imm interface does not work for me (and BTW cant find it on the CLI IMM) - documentation refers to certain conditions needing to be met, but does not explain what these are - maybe its a bug, maybe its something im doing wrong, so hoping your documentation people are planning on adding some more detail on this feature. i had to go down the path of getting someone at remote site to use F1 on console to set BootOption.BootOption to "PXE Network=Hard Disk 0", build off PXE, then reboot with the DHCP and tftp server disabled, wait about 10 minutes for PXE to try 12 or so times on each network interface before proceeding to the hard disk boot, before being able to fix BootOrder with asu - this is because I can't for the life of me find where you can modify the uefi settings via the IMM.

and also - while i think of it - is there a way to cut back the access rights of asu and MegaCli to just readonly? im worried that a mischeivous individual on a compromised system can do things like 'reset to factory' on the raid controller or uefi - it seems the only option open to me currently is to disable the USB interconnect to IMM completely, but that doesnt help with the RAID.

otherwise - love your work. keep it up...

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

webadmin@intel.com — Tue, 15 Sep 2009 20:22:51 GMT

Thanks for your feedback. I woul like to know your boot performance when we drop our Oct code drop.

Intel Communities: Message List - IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Firmware TypeVersion StringRelease DateIMMYUOOC7E09/30/2011UEFID6E154A09/23/2011DSADSYT89P10/28/2011

I must say I am sorely dissapointed with the "speed" of USB booting in the legacy BIOS mode in the IBM UEFI implementation.

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Firmware levels should be at:BIOS 1.04, Build P9E130AUSDiagnostics 1.13, Build P9YT40ABlade Sys Mgmt Processor 1.05Qlogic QMI2572, BIOS Revision 2.08IMM (Integrated Management Module) 1.05, Build YU0032F

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Re: IBM x3550 M2, x3650 M2, HS22, and dx360

Firmware Type Version String Release Date
IMM YUOOC7E 09/30/2011
UEFI D6E154A 09/23/2011
DSA DSYT89P 10/28/2011

Firmware levels should be at:
BIOS 1.04, Build P9E130AUS
Diagnostics 1.13, Build P9YT40A
Blade Sys Mgmt Processor 1.05
Qlogic QMI2572, BIOS Revision 2.08
IMM (Integrated Management Module) 1.05, Build YU0032F