MMC + dual W5580 + services = broken?

[db0x0]

I've run into an interesting problem that I haven't run across before. I also haven't been able to find any reference to it online after a few days of searching.

Our company uses a medium size in-house renderfarm for doing 3D animations. Most of our machines are rackmount systems, with several workstations rendering in off-hours (total of about 30 machines)

All are running XP64, most are dual Intel E54XX series Xeon processors, so 8-core systems. For years, I've been using the microsoft management console with a services snap-in for each machine to control the various services and log-ins that go along with our rendering software and have never had any problems (either now or back when we sere just vanilla XP)

In the last few weeks we've added a few dual W5580 Xeon machines. With hyper threading these show up as 16 core (thread) machines. In setting them up, I go through my usual procedure of adding them to the list of my msc file. When I open the MMC to access these new machines services, the first click on the machine will show the services list as normal, but a lot of the descriptions are missing and there are no stop/start services options. If I click on the properties of any of the services I get an 'unable to query status' error.

I click off of that machine in the list, then click back on it again I get an 'invalid handle' error, and can't see any info on the machine at all until I close and re-open the MMC again. (but the other non-W5580 machines work fine the whole time)

Now the odd parts. If I just disable multi-processor support in the machines BIOS, it all works perfectly. If I just disable hyper threading it all works perfectly. Disabling both and it works perfectly. Turning on multi-proc + hyper threading and MMC gets cranky again.

Also only affects networked machines. My local workstation is one of the dual w5580 machines, and can operate the services via MMC just fine as long as I'm local. If I work on a different machine, my regular workstation shows the same symptoms.

It's a very odd problem that I haven't run into before. Any ideas?

SJ

[MANEMAN]

Hi there, and welcome.

Try pulling the plug (Disconnect from mains supply) on all 30+ machines, and leave them off for about an hour.

Re-boot, and see if the problem is solved. Rather basic, but what works for the humble "Sky Plus" (Tivo for you in the U.S) box can sometimes work just as well for more complex setups.

You have already spent several days of searching online, so 1 hour of downtime is worth a shot.

If you have an I.T department try to get together with other members using the logic that two or more heads are better than one. (i.e - dont try to squirrel. :) )

If it is swallowing too many man hours, think about farming it out to a Tech subcontracting team. Many times it can be a more economically viable proposition. Often we can get bogged down, and need someone with a different viewpoint who can think "Outside of the box".

I think I am the only one still up at this hour, ( 2.30am, and a Migraine to die for !) so do please wait a day or two in case anyone else here comes up with a suggestion.

John.

PS. Am I the only one who thinks of "Interesting Problem" as an Oxymoron ? - Perhaps it's just age.

Which part of the U.S. are you from ?

[db0x0]

I guess interesting in the fact that I've never had a computer issue that you couldn't get a pile of Google hits on with just a few minutes of searching. Whether somebody found a solution or not is another matter, but very rare that somebody else hasn't had a similar issue on some forum somewhere on the entire internets.

For a simple power-down and wait cycle, that doesn't seem to affect it. A couple of the machines have been moved around since being set-up, generally being off for at least an hour if not a day or so.

We have an IT, but probably won't get a ton of sympathy since technically there's nothing wrong with the computers (they still render and work fine)

It's just more of an inconvenience that I have to remotely log into individual machines instead of doing everything from just a concise MMC

We're in the pacific NW, Seattle area.

SJ

[MANEMAN]

Thanks for coming back to us.

For many years I was a T.V / Audio engineer. Most nights I would sit down to eat my evening meal, the phone would ring and some B********** would tell me that they had got "An Interesting Problem".

From then to now I have adopted "Interesting Problem" as a personal Oxymoron. :) It wasn't meant as a dig. ;)

In electronics and computing you get these "One Off" situations arising, which is probably why you cannot find any reference to your particular "One Off"

Diagnosticians in any field tend to favour a previously tried line of enquiry. When you get something crop up like you have, whereabouts along a line of enquiry do you start ?

Remote diagnostics make it even more difficult (EMail,Forums,etc;) because there can be no physical contact with faulty equipment. Complex diagnostics via the internet often turns into a shambolic game of ** Chinese Whispers ** People start getting tetchy with each other, and very soon if you dont take control the whole thing can turn into a Flame War.

When an error has occurred more than three or four times it falls into the category of "Stock Fault". Lines of enquiry have been made, diagnostics have been started, - the ball is rolling.

The answers then come more easily.

You seem very adept at explaining things in a clear manner. Try other areas to petition your question.

Try that I.T. section. - You've got nothing to lose.

John.

[MANEMAN]

Had to come back to you again. :)

If you are gonna sort this one out on your ownsome.

This guy may just give you the impetus to start gaining on a more lateral way of thinking.

** Rajesh Setty - Life Beyond Code **

John.

[db0x0]

no fixes yet, but I can confirm I have the same issue with 4 new dual X5570 systems as well.

SJ

[MANEMAN]

Could be that the problem is starting to drift into the area of "Stock Fault" very soon, and you will be able to off-load it onto the I.T. Dept.

Good news from your point of view. :)

John.