Warning: Intel Server System R1304BTLSHBN (1U Hot Swap) Has A Drive Height Restriction

Due to the fact that the new Intel E3-1200 series pedestal and 1U server configurations are a little restrictive, we have been reserving their use for standalone SBS 2011 OS installs (OS on physical box).

Almost all of our server deploys are configured with either 2.5” 15K SAS, 3.5” 15K SAS, or Intel X25-M or 320 Series SSDs.

There is an Intel Technical Advisory dated May 6th, 2011 that states that the new Intel Server System R1304BTLSHBN 1U with four hot swap drive bays has a drive height restriction of 25.9mm.

Well, we just checked Seagate’s Web site for the height of the 300GB 15K SAS drive and found that its specification is 26.10mm.

Here is a look at the hard drive compatibility list for this unit:

There seems to be a distinct lack of 3.5” 15K SAS drives on the above list. We have a few 300GB 15K Seagate SAS drives kicking about for lab purposes so we would have been testing this chassis with them and discovered this issue right away.

It is good to know ahead of time that there may be compatibility problems with our standard configurations on new platforms.

The above page can be found via the following link:

• Intel Server Products: Tested Hardware List

Note that the above link jumps into the base THOL page for all of the new E3-1200 series products:

It looks as though Intel is moving away from a PDF, and thus offline, version of the Tested Hardware and Operating Systems List (THOL which is our definitive compatibility guide) towards an online version only:

Hopefully this new version of the Server Configurator has its stability issues dealt with. So far, the site seems to be functioning a lot better than it did just a few short weeks ago.

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

UPDATE: Stock on Intel Server Board S3420GPRX I/O Shield

We have found out that D&H USA has stock on the Intel Server Board S3420GPRX I/O shield (AGPRXIO).

Unfortunately D&H Canada is not allowed to sell Intel product into Canada at this time, so we were not able to get stock through their Canadian distribution point while the US distribution point will not drop-ship into Canada.

Fellow MVP Russ Grover suggested we talk to Applied Computer Online Services (Link to the I/O Shield part). They show available stock in their online ordering system.

But, because we are in Canada we ended up needing to call in to work out the product cost and the cost to ship up to Canada. In the end they were very helpful and their price was reasonable for each unit including shipping.

Our previous blog post:

• MPECS Inc. Blog: Setting Up an Intel S3420GPRX Entry Level Server Board In An Intel Pedestal Chassis

Note that we specifically requested FedEx as the courier since UPS charges an arm and a leg for their customs clearing services (we subsequently contracted the FedEx folks due to ongoing problems with UPS).

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Setting Up an Intel S3420GPRX Entry Level Server Board In An Intel Pedestal Chassis

We are building out a number of pedestal based servers over the next couple of months that will be standalone Hyper-V hosts for SBS 2011 Standard, at least a Remote Desktop Services server, at least one Windows Desktop OS, and in a few cases a dedicated BlackBerry (BESx) server.

The perfect platform for this setup is the Intel Server Board S3420GPRX (Intel Ark Site) as it has four Intel Virtualization Technology accelerated Gigabit NICs that can be teamed together along with a fifth Gigabit NIC that can be bound to an Intel Remote Management Module 3 (Intel Product Site) for external Web based console access.

The catch with this setup though is that the S3420GPRX is designed to be installed in a 1U form factor chassis. So, the retail box comes with _no_ I/O shield for the back of the pedestal chassis.

However, we are able to order them . . . hopefully:

• Intel S3420GPRX I/O Shield: AGPRXIO

We will see what our distribution channels say as we have put out inquiries to all of them and we have a request into our Intel sales representative.

So far, no one has indicated the what/where/when with regards to the parts. But, we remain hopeful.

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Not All UPSs Are Created Equal – Surge Let-Through

As Chris Knight pointed out in a comment on our post on a failed APC BR1500LCD and both power supplies in an SC5295BRP chassis, the Surge Let-Through value is a very important one when it comes to determining what UPS needs to be put in front of expensive and critical IT equipment.

So, here we have three APC UPS products set up in an online comparison on APC’s site:

APC Smart-UPS 1000VA SMT1000RM2U $559.00

APC Smart-UPS 1500VA SUA1500RM2U $719.00

APC Smart-UPS 1500VA SMX1500RM2U $779.00

Now, out of the above units, which unit or units would have a better Surge Let-Through number?

The answer was actually a bit surprising:

APC Smart-UPS 1000VA SMT1000RM2U 0.3%

APC Smart-UPS 1500VA SUA1500RM2U 0.3%

APC Smart-UPS 1500VA SMX1500RM2U 5.0%

What caught us off guard was the fact that it was the most expensive unit that had the lesser surge let-through rating.

Now, considering that the BR1500LCD unit also shares in the 5% rating, if we had the SMX1500RM2U in its place we probably still would have had two blown power supplies.

Thanks to Chris, we will be paying a lot more attention to that rating on any of the UPSs we supply our clients or purchase for our own use.

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Integrating An Intel RAID RS2BL040 Battery Backup AXXRSBBU7 With An Intel Server System SR1630HGPRX

The Intel Server System SR1630HGPRX has a spot on one side of the chassis to mount the Intel RAID Controller’s RAID Smart Battery.

Having the battery unit mounted elsewhere in the chassis is a good idea since having it mounted on the RAID controller itself causes an obstruction to the air flowing towards the RAID controller’s processor heatsink.

Tray Mounting

The plastic tray was a bit of a bear to mount in place:

We had to slip the guides in onto the high points of the mounting slot and somehow pivot it into the low point with a good amount of effort required to get the locking mechanism to snap in place. The lock is just above the white plastic zip tie in the above screenshot.

Note the star shaped pin near the bottom right hand side of the tray. One would initially gather that it was meant to guide the battery backup’s PCB board into the proper place. Especially since there looks to be a hole in the PCB board just for that purpose.

In fact, it is not a guide but a stand-off to prevent the PCB board from being pushed too far into the tray. Keep this in mind when mounting the battery and PCB.

Another catch to getting the battery/PCB in is in fact the catches themselves. There is one on either side of the tray’s lock. They provide a significant amount of resistance to getting the PCB to set in place. We needed to be mindful of just how much they were being pulled back while trying to slip the PCB into place.

And one more thing when it comes to the tray and batter/PCB combination, there is a requirement for the battery to sit above the PCB in order for the PCB to catch under the two mounting clips on the tray. The PCB clip that holds the battery down looks large enough to prevent this, but there is enough flex in the PCB itself to get things into place.

All in all, this is one snug fitting trio.

Remote PCB and RAID Controller

Once we ran the cable from the battery backup to the RAID controller, we needed to install a small daughter card onto the RAID controller itself.

The daughter card, shown above, has the socket that connects to the RAID controller as well as the black pivot pin that is used to set the mounting screw into on the RAID controller. The pivot pin is shown in the correct location above. We needed to loosen the screw on the daughter card and rotate the pivot pin into the correct location prior to installing it onto the RAID controller.

The daughter card installed:

After installing the power cable the RAID controller was connected to the PCI-E riser and finally installed in the system.

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Intel Resellers – Intel Real Server Hunt Contest Started This Week

We have entered these contests before as Intel runs them relatively frequently and they tend to hold a few “Oh, I did not know that” kind of gems.

• Intel Real Server Hunt Contest

The contest allows us to brush up on some key Intel server product features as well.

And, perhaps we will come away with some SWAG in the process. :)

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Server’s Having A Bad Day When … CLOCK_WATCHDOG_TIMEOUT (101)

We lose it almost continually throughout the day.

This particular box is our first Intel SR1630HGP Server System with an Intel Xeon Processor 3460 and 16GB of RAM.

This is what the %windir%\MiniDump directory looks like right now:

Our handy Crash Analyzer Wizard that comes with our Software Assurance and MDOP benefits tell us:

Click on that Details button and we see:

CLOCK_WATCHDOG_TIMEOUT (101)

An expected clock interrupt was not received on a secondary processor in an MP system within the allocated interval. This indicates that the specified processor is hung and not processing interrupts.
Arguments:
Arg1: 0000000000000019, Clock interrupt time out interval in nominal clock ticks.
Arg2: 0000000000000000, 0.
Arg3: fffff88001e5d180, The PRCB address of the hung processor.
Arg4: 0000000000000002, 0.

Now, it’s not like we have not seen this one before:

• Hyper-V On Nehalem CPUs Error – CLOCK_WATCHDOG_TIMEOUT Critical Hotfix
• Hyper-V CLOCK_WATCHDOG_TIMEOUT Error Within 24 Hours On Fresh H-V 2K8 R2
• This particular server hit by the freezes was actually the one from the above posts. It has been sitting off on a shelf as we have not had the need for it up until this current project we are working on.

We did run the hotfix on the server a while back, but apparently that version was not good enough since we brought all off the server’s server board firmware and the RAID controller’s firmware up to date. We did this just before bringing it back online.

The above screenshot of the MiniDump directory does not do the situation justice as the server has frozen a lot more times than that . . . leading us to look at other options for the setup we need.

But then, apparently the KB975530 hotfix has received a newer version from what we found when searching to find any further info on the problem.

Here is a screenshot of the hotfix directory with our original update and the one that we just received:

There was not a lot of time between v3 and v4 with the time of this post being late March.

It now remains to be seen if this fourth iteration of the patch will actually straighten out the timing problem happening with the Nehalem CPUs.

The server:

• Intel Server System SR1630HGP 1U
• Intel Xeon Processor X3460
• Intel RAID Controller RS2BL080
• Battery Backup Unit
• 16GB Kingston ECC DDR3 RAM
• 450GB 15K Seagate SAS drives in RAID 5
• DVDROM optical drive

***

This post has been sitting open for the last couple of hours while busy with other things. After consistently freezing today, the server has been up and running with the two server VMs and five desktop VMs without a hiccup.

Hopefully that will be the situation from now on!

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Hyper-V CLOCK_WATCHDOG_TIMEOUT Error Within 24 Hours On Fresh H-V 2K8 R2

We ran a test on one of our freshly installed Hyper-V Server 2008 R2 servers to see if we ran into this issue.

• Intel Server System SR1630HGP 1U
• Intel Server Board S3420GPLC integrated
• Intel Xeon Processor X3450
• 16GB Kingston ECC
• Intel RS2BL080 6Gbps SAS 2nd Generation PCI-E
• 3x 450GB Seagate 15K RPM SAS in RAID 5
• Intel RAID Smart Battery for the RAID controller

The server has only been up and running about 24 hours. We have been running some stress tests along with some test VM installs and backup restores as Hyper-V guests.

This is the Crash Analyzer Wizard’s results (DaRT MDOP product):

So, after a relatively short period of use the problem has reared its head.

• Our previous post: Hyper-V On Nehalem CPUs Error – CLOCK_WATCHDOG_TIMEOUT Critical Hotfix.

It looks as though the hotfix is an absolute must for any Technician’s Thumb Drive and a required install on any server that will utilize Hyper-V on Intel Xeon Processor 3400 and 5500 series CPUs!

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Hyper-V On Nehalem CPUs Error – CLOCK_WATCHDOG_TIMEOUT Critical Hotfix

Via Slashdot:

• Slashdot: Hardware: Microsoft Advice Against Nehalem Xeons Snuffed Out.

The Microsoft Knowledgebase article with the hotfix:

• MS Knowledgebase 975530: Stop error message on an Intel Xeon 5500 series processor-based computer that is running Windows Server 2008 R2 and that has the Hyper-V role installed: "0x00000101 - CLOCK_WATCHDOG_TIMEOUT"

We are just in the process of implementing Hyper-V Server 2008 R2 on the new Intel Nehalem architecture in both the Intel Xeon Processor 3400 and 5500 series.

So, finding out about this particular problem before the boxes went into production is a good thing.

The update begs the question though, does the hotfix really resolve the conflict with the CPU architecture?

It looks as though we will be doing a bit more testing than we normally do on new platforms to make sure that things do not blow up on us with the new soon to be deployed production Hyper-V boxes.

One of the linked articles from Slashdot is a good read:

• The H: Processor Whispers - About War and Peace

UPDATE: Here it is:

As a test, we did not install the update on a freshly installed Hyper-V Server 2008 R2 box and the above happened about midway through our second VM setup.

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Hyper-V On Nehalem CPUs Error – CLOCK_WATCHDOG_TIMEOUT Critical Hotfix

Via Slashdot:

• Slashdot: Hardware: Microsoft Advice Against Nehalem Xeons Snuffed Out.

The Microsoft Knowledgebase article with the hotfix:

• MS Knowledgebase 975530: Stop error message on an Intel Xeon 5500 series processor-based computer that is running Windows Server 2008 R2 and that has the Hyper-V role installed: "0x00000101 - CLOCK_WATCHDOG_TIMEOUT"

We are just in the process of implementing Hyper-V Server 2008 R2 on the new Intel Nehalem architecture in both the Intel Xeon Processor 3400 and 5500 series.

So, finding out about this particular problem before the boxes went into production is a good thing.

The update begs the question though, does the hotfix really resolve the conflict with the CPU architecture?

It looks as though we will be doing a bit more testing than we normally do on new platforms to make sure that things do not blow up on us with the new soon to be deployed production Hyper-V boxes.

One of the linked articles from Slashdot is a good read:

• The H: Processor Whispers - About War and Peace

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

Intel Memory Configuration Tool – Relieve That Pain!

One of the most significant brain busters with Intel’s new Core i7 architecture as well as the new Intel Xeon Processor 3400 Series and 5500 Series architectures is trying to get a straight answer on how to come out with a certain amount of RAM in the system.

This is the Intel Server Board S3420GP Xeon 3400 series memory page out of the product’s technical product specifications manual:

There are more pages both before and after this one that get into how to configure memory for this particular server board without really making things entirely clear.

There is the Intel Server Configurator Tool that will only allow certain RAM part combinations during the configuration process, but it does not tell us why it allows or does not allow us to choose a certain amount of RAM like 24GB for one box.

In comes Intel’s new Memory Configurator Tool:

We are looking to have a Xeon 3400 series server configured with 24GB. This tool will actually show us the how/what/where/why of getting there.

• Intel Memory Configuration Tool Download 

The tool at least takes some of the guess work out of configuring servers.

The above 24GB configuration request yielded the following result:

It says we can do 6x 4GB Registered DIMMs at 800MHz up top with the bottom right quadrant indicating the DIMMs can be Single, Dual, or Quad Rank when it comes to the memory chips on the DIMMs themselves.

Now, is the tool right?

Well, we will find out in short order since we have an SR1630HGP 1U server system based on the S3420GPLC series server board with six 4GB Kingston quad ranked registered ECC RAM sitting in the shop waiting to be built.

When we run through the Intel Server Configurator Tool and come to the memory configuration portion, we get the following:

The Configurator indicates up to 4x 4GB Registered Quad Rank DIMMs which does not meet our 24GB need. So, we shall see soon enough which one is right.

Philip Elder
MPECS Inc.
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer