The collected works of recondas - Page 12

viewtopic.php?f=1&t=16721811&
viewtopic.php?f=3&t=16720118&p=7292700&#p7292700
viewtopic.php?f=5&t=16723061&p=7319830&#p7319830

weldelandia wrote: Hi, is there a way to find users by location?

That's a fairly tall order. Not only can users put just about anything in the location field of their user profile, entering a location isn't mandatory.

You could always google the forum by the geographic entity of choice, something like "site:nekochan.net finland". If you're interested in finding members in a specific area, you'd probably get better results by a post that outlines who you want to find <and your reasons for doing so>.

3dchris wrote: Also, I don't see the option to turn on "Smooth Wireframe", or "High Quality Viewport". Are these options not supported on the SGI?

Perhaps not what you're looking for, but the Maya 6.5 Help Menu mentions the following:

Maya 6.5 Help Menu wrote: Shading > Smooth Wireframe

Displays smooth wireframed objects in Maya's 3D views, including the Hardware Render Buffer and the 3D Paint Effects view.

Note: Does not work in 2D views, including the UV Texture Editor.

<and>

Shading > High Quality Rendering

When high quality interactive shading is turned on, the scene views are drawn in high quality by the hardware renderer. This lets you see a very good representation of the look of the final render without having to software render the scene.

87Porsche wrote:

recondas wrote: Yeah - its almost enough to revive speculation that an EPS power supply from a PC could be rewired to work <maybe even a Sparkle >. If you have the chance a few close up photos of the label, interior and any makings on the PCB would be very interesting.

Here you go. I thought I had thrown it away.

The power supply from your Fuel is labled as a Sparkle FSP460-60PFN. Sparkle/FSP Group has a data sheet for the FSP460-60PFN power supply used in your Fuel / shown in your photo. A cursory google search shows it sold new for $70 to $80, and there's currently one on eBay at $25: http://cgi.ebay.com/FSP460-60PFN-460W-A ... 45f87284e1

No longer a stocked item, but <minus the part number tags added by SGI> the photo of the label of the Sparkle FSP460-60PFN on NewEgg's web site appears to be nearly identical to the label in your photo: 87Porsche, could you post a photo of the the connectors that attach to the Fuel motherboard? SGI might have used a non-standard wiring arrangement in the connector, but other than that I don't see anything that would suggest that the SGI version isn't identical to the generally available PC version.

No guarantees, but it might provide a reasonably priced route to revive your Fuel.

zski128 wrote: What was the reason for the later higher wattage PSU? Higher CPU speeds, more expansion cards?

I don't know that SGI has ever publicly given a reason, but your list of suggestions sounds like it includes most of the usual suspects.

Black Cardinal wrote: Forum question: How do I edit the thread title to match the new configuration?

I don't get quite the same edit panel you do, but to the best of my pre-moderator recollection the title changing bits were the same. You'll have to edit the initial post to change the title of the thread.

In the top left panel of the 'Edit post' window should be selections for "Delete post", "Post icon", "Subject" and "Message body". The field to the right of "Subject" should be a editable text box containing the thread title. Changing the contents for that text box will change/update the thread title.

From a conceptual standpoint the wiki makes better sense - articles are more focused/on point than the ad-hoc entry style of a forum. Having said that, the formal presentation style of the wiki doesn't readily foster the level of social interaction that's found in the laid-back atmosphere of the forums - which might partially explain why the wiki has been slower to gain acceptance with some of the regulars.

zafunk wrote: Wikify the aggregators and keep the individual threads going for discussion purposes.

So for that reason I'd agree with zafunk's suggestion to keep a forum discussion in addition to wiki articles would be a good idea - it'll keep another route for input from anyone who might not know the in-n-outs of the wiki interface <or doesn't find the level of wiki peer interaction as socially stimulating>.

For now I'll split the tape drive conversation out to the general hardware forum. Some of that info may prove helpful to someone with an HP DAT drive <with an O2 or any other SGI/MIPS system with a SCSI interface>.

HP published a Unix Configuration Manual that included a IRIX kernel configuration file for their LTO/Ultrium drives. The trick is/was finding a copy. The current <commonly available> version is Edition 5, from which HP unfortunately dropped IRIX.

After playing word games with google for a while I managed to find a copy of Edition 1 of the HP Ultrium Tape Drives Unix Configuration Guide - which does indeed include an IRIX 6.5 kernel configuration guide <and a few other gone-with-the-wind unix flavors>.

Seeing as it's an Edition 1 manual, it probably deals with earlier model HP LTO/Ultrium drives. The manual is still publicly available on one of HP's semi-obscure servers: http://www.google.com/url?sa=t&source=w ... 1g&cad=rja
As long as it's still available I'd recommend downloading a copy - HP gives some advice on user changeable block sizes to tune the the drive for specific uses, as well as a run down on specific commands to use

Here's the kernel configuration file <watch for erroneous word wrap>:
For anyone who's not certain what to do with the kernel config file, there's a neko-wiki that gives the run-down on where to put the file and how to get the kernel to recognize it .

Nicely done fu!

BTW - after everyone has had a a day or two to catch up with the wiki-related posts here in the O2 Hardware Aggregator, I'll be splitting them off to join the New Nekochan Wiki thread. That will keep the O2 Hardware Aggregator on topic, and for long-term storage/posterity they'll be a better fit in the neko-wiki topic.

hirudin wrote: I know moving one of these can be awesome (and by awesome I mean a pain).

Just in case anyone outside of LA might be tempted - moving a single-rack system is pretty straightforward if you have access to a pick up truck and a helper. I picked up the one I had using this method , and it went to the new owner the same way <in a mini-sized pickup truck>.

dc_v01 used the same method when he picked up his Onyx2. He documented the process with photos here: viewtopic.php?f=3&t=16721076

As a follow up, I had a couple of Dell (oem'd HP) LTO-1 drives on-hand, and moved one to an external case to test compatibility with the HP provided kernel configuration file.

After the initial power up, as expected the drive appeared as "unknown" to both an hinv -vm and "mt status". The good news is that the drive has standard HP firmware, so no modifications will be needed to the device identification string in the HP provided kernel configuration file.

After adding the HP kernel config file to /var/sysgen/master.d/scsi I updated the kernel <with autoconfig> and rebooted. After the reboot, the drive was correctly recognized by an hinv and mt status:


After using a 75 GB DLT drive, the 200GB capacity of the Ultrium/LTO-1 is nice. I've got a full system backup running now - I'll update this post when its done.

Update: I used the IRIX provided GUI <for cpio> to run a full system back up - and after the backup completed was able to read the contents of the tape and restore selected files.

The LTO-1 drive was connected to a LSI 21320 Ultra320 controller <according to the hinv as an LVD device>. The full system backup was 60.71 GB. From start to finish it took 33 minutes, 30 seconds - a significant improvement considering the DEC DLT7000 35/70GB drive < the same one used when I wrote the tape drive wiki > needed several hours/usually overnight to make a similar sized backup of back up of my Octane <connected to the Octane's native single-ended UltraSCSI bus>.

As long as we're dealing with SGI hardware of roughly the same era <O2, Octane, O2k, Onyx2, IP35, attributing specific SCSI devices to individual hardware platforms might be a little too fine-grained <SCSI devices such as hard drives, DVDs, CD-ROMs, CDRWs, but *not* PCI SCSI controllers> . I've worked my way through pretty much all of those systems, and can't recall a SCSI device that worked on one but not the others. I might have just been lucky, but if not, it'd save quite a bit of wiki creation time.

If a specific device appears in only one or two of the forum aggregators, it would be simple enough to add a note mentioning "not yet tested on platform X".

Ditto for third-party monitors, and a possibly a few other categories as well.

Here are a few more kernel configuration files for tape drives that are larger/newer than those included in with IRIX. These are compliments of a web search and I don't have the specific tape drives to test with, so no guarantees.

In an attempt to consolidate, the last four were recopied from videobrat's 2005 nekochan post .

In a follow up to videobrat's original post, LoWeN mentions that an additional piece of <licensed> IRIX software is required to allow the use of tape drives larger than a DLT IV. The additional software mentioned by LoWeN is an asynchronous personality daemon, or APD. Here's a forum post by videobrat where he describes problems with a Quantum LTO-3 drive used without APD - and a follow up post where he mentions that an HP LTO-3 drive he got as a replacement for the Quantum LTO-3 worked without APD .

VideoBrat's initial problems with the Quantum LTO-3 drive may have been the result of a typographical error in the kernel configuration file he used, rather than the lack of APD software. The first line of the IRIX kernel config file he used read:

{ DATTAPE, TPDAT, 8, 7 , "CERTANCE", " ULTRIUM 3 ", 0, 0, {0},

- the fourth field in that line "7", should reflect the total number of spaces in the fifth field ("ULTRIUM 3"), which is actually 9 spaces. To comply with SGI's guidelines, it should have read

{ DATTAPE, TPDAT, 8, 9 , "CERTANCE", " ULTRIUM 3 ", 0, 0, {0},

(the copy quoted below is the edited version).

I don't have the APD software, and so far that hasn't been a problem with backups and partial restores using the non-APD HP LTO-1 I tested with. In the absence of the APD software, the best practice may be to test a non-APD backup *before* a critical system back up is done. ( UPDATE : the list of successfully tested drives larger than DLT IV now includes an IBM LTO-2, a Quantum (Certance) LTO-2HH, and a Quantum SDLT320 - follow the thread for additional details).

There's not much in TechPubs on APD beyond some <similar> customer letters < scroll down to see the list >, and an indirect mention in the IRIX 6.5.30 Release Announcement that certain drives <that need the APD software> would appear in "mt status" as an "Async Personality Device".

I suspect that <among other things> the IRIX asynchronous personality daemon *might* provide for faster read/write performance, but was also intended as a marketing device to guide customers contemplating the purchase of a tape drive back to SGI. One source that might shed additional light on the subject is Sony's Unix Configuration Guide v2.24 . In that manual Sony recommends using APD v2.7 for their AIT-3 and SAIT-1 drives, but also provides a IRIX kernel configuration file to allow "testing without APD" <perhaps Sony's way of backdooring APD when Sony's potential customers balked at having to purchase an APD license just to test Sony's drives> . Sony's inclusion of a 'test' kernel driver at least suggests some drives will function without the addition of the APD software, and might hint that not using APD doesn't result in a performance hit - it would hardly behoove Sony to provide a method to test their drives that resulted degraded performance. In either case, the ability to use any of the drives listed below could still offer backup capacities greater than DDS4 or DLT IV.

So if anyone does try one and have success, please post a follow up note.
HP Ultrium 1 / LTO-1
IBM LTO2
IBM LTO3
Quantum SuperDLT-1
HP Ultrium-3/LTO-3 <originally provided by videobrat in this nekochan post >
Quantum (Certance) LTO-2
Quantum (Certance) LTO3
HP LTO-4 (Model 1760) - provided by hhoffman in this nekochan topic.
The following were also provided by videobrat < in this original post >:
Quantum SDLT 220
Quantum SDLT 320
Quantum SDLT 600
Sony DTF-2
If any one has additional <or alternative> examples, please post them.

The voltage/color values are the same on the label of the SGI FSP460-60PFN Sparkle PS and the Sparkle PC version of the power supply (with the same FSP460-60PFN part number). Having a similar photo of the connector and wiring bundle of the Sparkle PC power supply might help resolve what <if any> changes SGI made in the motherboard connector pinout.

If there aren't any visible changes, I'd say there's at least a reasonable chance the Sparkle PC unit might be a direct replacement of the (currently) hard to find Fuel power supply.

If they the voltage/color values in the connector bundle differ, then a few minutes spent with a DMM, a Fuel <and a working Sparkle power supply> might make re-wiring the connector bundle in the PC version a possibility.

EDIT : If anyone feels particularly brave <or otherwise desperate to find a reasonably priced Fuel PS>, an eBay seller currently has ten of the non-SGI Sparkle FSP460-60PFN power supplies listed for $49.99 (US) - with a 'make an offer' option: http://cgi.ebay.com/Sparkle-FSP460-60PF ... 5642a756a4 Check for yourself, but the connector doesn't appear to be wired to the same color-code combination as 87Porsche's Fuel example, so some work with a DMM is probably still in order.

I voted multi-processor because it represents all the SGI stuff I have running - a 16p Onyx300, a 8p Onyx O350 InfintePerformannce, and a 4p Tezro.

I do still have two O2s, but those are just sitting on the shelf unused, balanced out by a dual-processor Octane (in the same condition).

With 22 processors involved, that's a fairly complex piece of hardware to troubleshoot. I think the GOBI would suggest that minimizing the number of components tested at one time would simplify the process (something akin to the process suggested in this discussion ).

To avoid later issues with the PROM having components re-installed in different locations, put a masking tape tag indicating the position of each board and CrayLink cable before disconnecting or removing them.

• Temporarily disconnect the CrayLink connections and Graphics module and test only the compute module with the MSC that doesn't get to 3F3F3F.
• Remove all XIO expansion boards (X-Town, FC, MSCSI, etc.)
• After that module is powered up freestanding, examine the nodeboard LEDs for error codes , then test the system with only the problem nodeboard (in N1).
  • if you have issues with the normal console try the diagnostic port on the MSC, and
  • if you can't get diagnostic info with only the problem nodeboard installed, move it to N2 and put a working node in N1.
• As j-j has suggested follow that error code(s) to try to isolate the problem.
• If the problem can't be eliminated while only the problem nodeboard is installed, try a different nodeboard (in the same slot).

Once you get a nodeboard in the problem slot to display a post-initialization heartbeat , add the other components back, one-at-a-time.

and please post a follow-up on your findings - details on the problem from someone with your experience and knowledge of the hardware is almost certain to be of future help some one (who has yet to acquire either).

Never used it, but it looks like TechPubs has some documentation on the subject:
InPerson User's Guide
InPerson Setup and Administration Guide

and the man page for InPerson 2.2.1 (for IRIX 6.5) lists minimum hardware requirements:

canavan wrote: Anyone? If no one knows, could somebody please measure his?

DM2s, DM3s and DM5s have become very affordable, but the LVDS cables used to connect the DM2/3 to the DM5 are still pretty rare (and likely to get even more so when Richtom1s limited supply is exhausted).

If someone with a set of the LVDS cables would take a few minutes to meter the pinout it would be a nice gesture to the nekochan community.

This one arrived as a (nearly) empty chassis - added a system board (along with just about every screw in the chassis), processor node board, V10, and the extra cooling. This particular O350 belongs to another nekochan member who asked me to do the graphics conversion.

fu wrote: new rocker in your place? congrats dave

Thanks - it was a good an opportunity to put together some parts that I had/just acquired. Since this would make my third O350 with graphics, I'm going to pass it one along to someone who'll give it more use.

fu wrote: what is a SNOWBALL device? funky name...

The DM3/xtdigvid digital video board. (tho I used to think the name was a play on my chances of actually finding one).

ramq wrote: Nice one.

Thanks - they do make a very nice graphics workstation. Functionally the hardware is the same as the rack mount version of the Tezro or an Onyx (350) InfinitePerformance system. Once you add a graphics board the PROM splash with display either "Tezro" or "Welcome to Onyx InfinitePerformance" (your choice): viewtopic.php?f=3&t=16721235&

ramq wrote: Since you have three of'em, did you follow the threads before regarding Vpro in O350 or did you come up with some new nifty solutions to make it just as close to a "factory" config as possible? I'm thinking cutting metal, replacing fans, etc...

There are two basic methods. In either case the VPro board will overlap the processor cooling shroud (which causes the VPRo heatsink to foul the O350 processor shroud). The major difference in the two methods is how clearance for the VPro heatsink is acquired.

Neither method requires making any permanent changes to the O350 - any O350 parts that are removed are undamaged and easily replaced (with a bolt or screw) if you wish to return the system to its original server configuration.

The first method requires less metal cutting, but will work only with two-processor O350s. This method requires the removal of the processor cooling shroud (pull the node board and remove the attachment screws from the bottom - but avoid the screws for the RAM supports), and one of the (2) radial exhaust fans (the one blocking the XIO-2 slot on the PCI bulkhead). Removing the processor shroud makes room for the VPro heatsink. The only metal cutting is trimming the sheet metal bulkhead on the VPro (a five minute job with a pair of tin snips), and cutting a small piece of user-supplied angle stock as a fan support. The reason this only works with two-processor O350 is the placement of cooling fans in the space the third and fourth processors would occupy in a four-processor O350. bplaa-yai posted some explanatory photos of this type of conversion: viewtopic.php?f=3&t=16719768&p=7303988&#p7303988

The second method will work with either two or four processor O350s, but requires trimming the rear of the VPro heatsink (because the O350 processor shroud is left in place) and installing a radial fan under the graphics board (in a similar fashion to how the factory versions are cooled). The VPro heatsink is easily removed from the PCB and is soft enough to be cut with hacksaw or powered metal saw. There's a write up with photos in this thread (scroll down past the hinv): viewtopic.php?f=14&t=16721099&

The graphics cooling fan(s) will attach to a previously unused three-pin fan header on the O350 system board, and SGI uses a non-standard connection for the fan wiring. Details are in this post: viewtopic.php?f=3&t=16719768&p=7301078&#p7301078

mapesdhs wrote: how right you are. I currently have the cables priced higher than the DM2.

Since the price probably limits your (potential) LVDS cable customers to those with corporate budgets - the very kind who wouldn't be caught dead mucking around with a soldering iron and therefore would still buy from you even if the cable pin-out was publicly available - does that mean you're graciously volunteering to help out your fellow nekochan members by taking five minutes to meter the LVDS cable pin-out.

Or were you just mocking our plight?

mapesdhs wrote: I'll do it when I can...

Thanks Ian - your offer of help is greatly appreciated.

fu wrote: guys would a snap of the cable suffice? or do you have to crack it open and meticulously examine its guts?

What's needed is a listing of how the wires in side the are routed from one end to the other. While a photo would help identify the cables in question, the needed detail would be hidden from a normal photograph. On the other hand, no disassembly should be required, the internal wiring can be determined from either end with the use of a multi-meter or even a simple continuity tester.

fu wrote: i'll be doing some post work next week, i can ask around if there's any left/hidden in the storage room

If there are any laying around unused, send 'em to me and I'll test them and post the pin-outs.

Rincewind wrote: What kind of sled are used in the O300?

Like they did with the O350, SGI used an off the shelf drive sled in the O300. The O350 sled was also used by Intel/Sun/others, the O300 sled in a few IBM servers.

The O300 sleds are marked with assembly number A06447-00x (where the "x" is either a 1 or a 2).

A very brief eBay search turned up a couple in Italy for 15 euros each: http://cgi.ebay.com/rack-hd-SCSI-SCA2-L ... 5633674311

Except for the latching mechanism, the sleds used by O350 and O300 are identical. If all else fails, you could modify the slightly more common O350 sleds to fit into an O300. bplaa-yai provided photos of the process in this post: viewtopic.php?f=3&t=16719820&

SAQ wrote: So SGI used "real" drive sleds as baffles on the O3x series?

Pretty much, like ianj mentioned, they used a commodity sled with a removable air baffle.

The removable baffle is the tan part in the photo. Hard to make out in my camera phone photo, but "Remove only when installing hard drive device Keep for future use" is printed in the bottom of the baffle in this O300-style sled:

smj wrote: I for one am very glad they went this way in the O3x0 units... I suppose the same holds true for the Altix 350 and 450s, no?

The Altix 350 / Prism used the same sled as the Origin 350. Haven't ever seen an A450, but the illustrations in TechPubs make it look like they also use the same sled as the O350: http://techpubs.sgi.com/library/tpl/cgi ... #id5202614

There are screen shot shots of several of IRIX apps in the gallery - perhaps a few you don't already have (and you could link the SS to the entry if you think it's appropriate).

Be sure to look in ajerimez's gallery - it's an especially nice resource for application screen shots.
http://www.nekochan.net/wiki/gallery2/v/guest/

Here's a few shots of completed fan installation. Installation of a VPro Graphics board requires removal of one of the two 90CFM radial exhaust fans (Pabst RER 101-36/12/12A) found in the headless server version of the Origin 350. With the loss of 90CFM of air flow the issue of primary concern when adding graphics to the relatively restrictive confines of the 2U-sized Origin 350 is cooling. The dual 90CFM fans flow enough air that a stock O350 will heat up if run with the case open, while an O350 with one of the exhaust fans removed (and a graphics board installed) will run cooler with the case open.

This Origin 350 graphics conversion used bplaa-yai's method of removing the processor shroud and installing three fans between the processor node and the VPro graphics board. This method only works with two-processor systems because the fans are positioned where the third and fourth processors would be in a four-processor system. If you have a four-processor Origin 350, or prefer not to remove the processor shroud from a two-processor system, an alternative method of adding VPro graphics is described here: viewtopic.php?f=14&t=16721099 (which is similar in concept to SGI's method of cooling a factory built Onyx InfinitePerformance). Neither method requires making an non-reversible changes to the base Origin 350.

The fan support is a piece of heavy gauge sheet metal angle stock .5"x.5" wide by 10.25" long (the piece of sheet metal angle stock was selected because it closely matches the appearance of rest of the O350 chassis). It's supported on the PCI-side of the O350 by one of the threaded tabs used to secure the case lid, and by a fabricated tab on the power supply side (the tab was added so that the fan support would bisect the chassis at a right angle). The mounting tab on the power supply side is a simple 1" long x .5" wide section of heavy sheet stock with two drilled holes - one end is captured under the screw that closes the power supply cable clamp, the the other end extends outward to support the new fan bracket. The hole drilled hole in the other end of the tab was sized to fit a surplus fan push-pin mount (used to secure the newly added fan bracket to the tab), but a screw and nut would work equally well. To ensure that no contact occurred between the added fan bracket and the power supply cable bundle, the power supply was moved to the outside position (the O350 chassis has provisions for dual power supplies). O350's with a single power supply configuration normally have the power supply installed in the inside position (nearest the node board). In this instance the single power supply was moved to the outside position to ensure the added fan bracket wouldn't make contact with the power supply cable bundle. To prevent any issues with environmental checks done by the L1 controller, the power supply cable bundle was left attached to the same header on the system board (the power supply cable bundle attaches to the power supply with an edge-connector, so the cable bundle can be disconnected at the power supply end and moved there).

The two fans on the outside of the tree-fan array are two 60mm x 25mm Vantec TF6025 Thermoflow temperature controlled fans. These run at between 1900-4200 RPM, 9.9-24.2 CFM and 19-33.5 decibels, and were selected to provide reasonable cooling with minimal additional noise The center fan of the tree-fan array is a 60mm x20mm SuperRed CHA6012CS-A that's rated at 3900 RPM, 19.4 CFM and 29 decibels. The tach wire on the SuperRed is the only one of the three that's wired to the tach pin on the O350 fan connector. The SuperRed was selected because of its static 3900 RPM speed rating. The L1 in the O350 will trigger a (non-fatal) fan warning if the fan attached the the three-pin graphics fan header (located on the system board) drops below 1679 RPM. During a cold start the L1 environmental controller lowers the fan speeds by a percentage of the overall speed rather than to a specific RPM level. That percentage of the 1900 RPM start up speed of the Vantec fans isn't high enough to avoid triggering the warning, so the trick is to select fans that have a speed rating high enough to meet that they'll be above the L1 graphics fan warning level (1679 RPM). The the start up percentage of the fixed 3900 RPM speed of the SuperRed fan is around 1800 RPM (enough to avoid the 1679 RPM L1 fan warning), and even after hours of constant use hasn't exceeded 2200 RPM. At those RPM levels the SuperRed fan isn't noticeable over the original equipment power supply fan or radial exhaust fans, so as an alternative, three SuperRed CHA0612CS-A fans (or a similar fixed speed fan) could be used in all three positions with out adding to the O350's noise footprint.

Regardless of the fans (or cooling method) selected, particular attention should be paid to the fan connection pin-out. The fan header on the O350 system board uses a non-standard pin-out: Pin 1 - 12VDC; Pin 2 - Tach; Pin 3 - Ground (the location of pin 1 is screened on the system board). A 4-pin PWM fan may also be used; the O350 fan header and the PWM fan socket are keyed to allow connection of only the 3 necessary fan wires (although the 12VDC, tach and ground wires should be configured to match the O350 pin-out).

Because the Bedrock ASIC is usually shows up in an L1 'env' query as the hottest part of the system, a low-profile variable-speed thermally-controlled 70mm CPU cooler was added to the top of the Bedrock heat sink (70mm is a perfect fit). The cooler is secured by double-sided foam tape placed between the fan frame and hub and the heatsink. This lowered the Bedrock temperature an average of 8C, and the fan wasn't acoustically noticeable in the running O350.

There's been a few questions in the past concerning support for the VPro PCB. Fuel VPro boards come with a sheet metal support rail screwed to the edge of the board directly opposite of the XIO2 edge connectors. As it comes out of the Fuel that support rail is longer than the VPro PCB by several inches, and as a result would protrude into one of the added cooling fans. The one in the photo has been trimmed so that it was the same overall length as the VPro PCB (using a cut-off wheel, but a hacksaw or even tip-snips would do the job).

maxsleg wrote: I am NOT impressed by Avis or HWY 101).......

Sorry to hear about your troubles - but hope the closure was transient. Be passing through there pretty soon on the bike on my way to take a right at Puget Sound.

smj wrote: (I've stuffed this info and a picture into the wiki for future reference.)

Thanks - here's a couple of photos of the part identification labels and the sleds side by side.

If you have more than one system of a particular type (and the hardware is similar enough to support the same base install), one possibility would be to install just one and make clones of that disk for the rest.

Here's a link to Ian's Cloning guide: http://www.sgidepot.co.uk/disksfiles.html#CLONE