SGI Octane System Module Stuck
Jeff Corbets
I just purchased a SGI Octane and received it yesterday. Before
powering it on I wanted to perform a good cleaning of the system and
tried to remove the system module following the instructions in the SGI
Octane Workstation Owner's Guide.
Unfortunately, after pulling the module out about two inches, it gets
stuck on something. I have removed the lockbar and all four memory
modules are seated on the board. I also have tried removing it with the
workstation laying on three of its sides with a decent amount of pulling
force. The module appears to be stuck on something close to the top of
the workstation, but it was not possible to see what it might be caught on.
If anyone has had a similar experience and can point me in the right
direction on how to get the module out, I would appreciate it.
Thanks,
Jeff Corbets
UNIX Museum
Jeff, are the 2 screwable lockbars (the ones on the module itself)
extracted all the way out? If one of them is pushed in the board, then
it will still lock it I think...
Jeff Corbets
the system module disconnected itself from the workstation. When I pull
on them, they do not come out anymore.
I'm guessing that something is slightly bent, I just don't have any clue
what the bent piece might be, or how to fix it.
Thanks,
Jeff Corbets
Gerhard Lenerz
> If anyone has had a similar experience and can point me in the right
> direction on how to get the module out, I would appreciate it.
Try putting it together again, reinsert the lockbar and start from
the beginning. On my Octane it seems that the locking mechanism somehow
gets stuck if the lockbar isn't removed properly (e.g. when I think
about pulling it too late and I had my hands already on the board
itself).
Gerhard
Jeff Corbets
and then taking it apart. The same thing happens - the module gets
stuck about two inches out. Is there some place where I can find good
instructions to take off the rear skin and somehow view or ensure the
locking mechanism has retracted?
Thanks,
Jeff Corbets
UNIX Museum
When you say "rear skin" you mean the main skin? It yes, then just press
on the front corners of the skins and pull up (or push them back). Once
they're off, you'll have access to the lock mechanism...
Jeff Corbets
the other piece - the main skin - is being quite finicky. The corners
to push on, are those as you view the machine from the front or from the
rear? Also, how much force do you need to use to push (slide?) it off?
Thanks,
Jeff Corbets
Jeff Corbets
the Octane Workstation Owner's Guide. Unfortunately, the locking
mechanism is working correctly, so now I'm back to square one. Should I
just pull out the system module as hard as I can, or is there something
else to try?
Thanks,
Jeff Corbets
Tony 'Nicoya' Mantler
wrote:
: Hey everyone,
As I recall, the some part of the shielding can occasionally get bent and jam
the system module when it's removed rougly about 2 inches from the case.
Luckily, SGI had the foresight to include a tool to deal with this: the lockbar.
Slide it into the gap between the system module and the case at whatever point
feels like it's sticking, it should be just thin enough and plenty long to reach
in and unhook the bent piece of shielding.
If all else fails, you should be able to remove the frontplane from the system
with the system module only partly extracted. The instructions on how to do so
(minus the hung-up system module) are available in the Octane Owner's Manual.
Don't forget to bend the shielding back into place when you're done extracting
the system module.
Cheers - Tony 'Nicoya' Mantler :)
--
Tony 'Nicoya' Mantler -- Master of Code-fu -- nic...@ubb.ca
-- http://nicoya.feline.pp.se/ -- http://www.ubb.ca/ --
barefoot
>Hey everyone,
I have had this happen to me a few times and each time it is because
I have not pulled out the plastic screwable lockbars properly.
Porperly=not too much and not little
Each time I have had to push the module back in jiggle the lock bars (each
individually) then pull it out until it seems to have disengaged then pull
the module out. the top lock bar seem to be more pedantic then the bottom
one (so if in doubt play with the top one).
I am rather experienced at doing this :-( as my dimms are always not
seated properly forcing me to take the system module out and re-seat the
dimms (yes I have cleaned the dimms, compressed air the slots .....).
Cheers
barefoot
Lewis Saunders
> As I recall, the some part of the shielding can occasionally get bent and
jam
> the system module when it's removed rougly about 2 inches from the case.
Yup, it's very probably that. Just to clarify, the things it gets stuck on
are at the top, and are flat strips of metal about half an inch wide which
go across just under the honeycomb metal grill. They're really hard to see,
but you can tell where they are from looking at the strips on the upper side
of the honeycomb. Try looking from the back through the gap above the system
module with a torch shining in the top. They get bent downwards,
unfortunately pulling the system module makes the problem worse...
If you can't unbend them with the lockbar, you can try pushing the module
all the way in, then pulling /really/ hard as fast as you can. After doing
this a couple of times I bent them so much that they weren't an obstruction.
Obviously this is not good for the board, but mine survived... build quality
:-)
Harry Goldsholl
is bent and ptreventing the Motherboard from being removed.
Pull it out as far as it can go and look inside with a flashlight for the
obstruction.
Jeff Corbets
enough to cause the system module to jam. Not thinking about the
lockbar, I ended up taking the whole machine apart, including the
frontplane, and then using the opposite end of a screwdriver extension
bit to push the bends back up.
Two questions: Is the only purpose of those flat metal strips to secure
the honeycomb if it somehow falls? Is there a reason why they get bent
down or is it a matter of creep in the metal over time?
Thanks everyone for their help. I took a few pictures of the bars and
the machine and will put up a small site in case other people run into
the same problem.
Regards,
Jeff Corbets
Greg Douglas
Jeff Corbets wrote:
> Yep, both of the metal bars that secure the shielding were bent down
> enough to cause the system module to jam. Not thinking about the
> lockbar, I ended up taking the whole machine apart, including the
> frontplane, and then using the opposite end of a screwdriver extension
> bit to push the bends back up.
>
> Two questions: Is the only purpose of those flat metal strips to secure
> the honeycomb if it somehow falls? Is there a reason why they get bent
> down or is it a matter of creep in the metal over time?
That's the only purpose of the metal strips, afaik. I've seen them
get bent down when people push on the top side of the honeycombs.
Unfortunately, the metal is one-way. Once it gets pushed down, the
ONLY way to get it back up is to push from the opposite direction,
namely inside the IP30 or gfx board opening.
>
> Thanks everyone for their help. I took a few pictures of the bars and
> the machine and will put up a small site in case other people run into
> the same problem.
There are 3 main reasons for not being able to remove an IP30 board
from an Octane. They are (in order of how common - btw, I've seen
them all):
A) Lockbar still in place (yes, I've yanked quite firmly, called
the Octane some names that would make a sailor blush, and then
discovered the lockbar. Hint - if someone is watching, tell
them you were just testing the mechanism.
B) DIMM modules fallen out of the DIMM slots, lying in the bottom
of the Octane chassis, keeping things from operating smoothly
(hint - set the Octane on its top side and re-insert the board)
C) The aforementioned bent metal bars on the honeycomb air filters.
(hint - use the lockbar as a prybar as Tony Mantler suggested).
--
Greg Douglas
Reputable Systems
http://www.reputable.com