20 May
2019
20 May
'19
7:39 a.m.
From: Jos Dreesen
And so the story continues....
https://ibms360.co.uk/
Wow, what a great recover, and a great site documenting it!
Renting temporary local storage was a great idea; it would have been hard
to get all that out of there on schedule any other way. (Alas, I don't
know any hauliers who can help them get it all back to the UK!)
They should have tried to get the raised floor too (I guess it wasn't
included in the sale), since i) it'll be useful if they try and get the
machine up and running, and ii) it'll probably just get scrapped. Although
there may already be raised floor where they're planning to put it.
Noel
20 May
20 May
8:38 a.m.
On Mon, May 20, 2019 at 10:39 AM Noel Chiappa via cctalk <
cctalk at classiccmp.org> wrote:
They should have tried to get the raised floor too (I
guess it wasn't
included in the sale), since i) it'll be useful if they try and get the
machine up and running, and ii) it'll probably just get scrapped. Although
there may already be raised floor where they're planning to put it.
I guess it would look right for the era, but I'd never build a data center
with raised flooring after my experiences with them. It's such a pain to
work with compared to a sealed concrete floor and overhead cable trays.
That said, I'd bet you can pick up raised flooring cheaply locally if you
find someone who works on commercial buildings. For that era, it's either
going to be wood or concrete core, and will cost more to get rid of than it
has in scrap value (between the dumpster rental and the injured backs
moving it).
Pat
9:10 a.m.
I guess it would look right for the era, but I'd
never build a data center
with raised flooring after my experiences with them. It's such a pain to
work with compared to a sealed concrete floor and overhead cable trays.
But with a raised floor, you can whack the tile puller down in such a
way that it makes an enormously loud pop that startles everyone in the
datacenter!
I cannot imagine difficult it would be to run S/360 era cables in
overhead trays. Many are an inch or more in diameter.
--
Will
--
Will
9:40 a.m.
On Mon, May 20, 2019 at 12:11 PM William Donzelli <wdonzelli at gmail.com>
wrote:
A good data center is loud enough that no one can hear that. :)
I guess it
would look right for the era, but I'd never build a data
center
with raised flooring after my experiences with
them. It's such a pain to
work with compared to a sealed concrete floor and overhead cable trays.
But with a raised floor, you can whack the tile puller down in such a
way that it makes an enormously loud pop that startles everyone in the
datacenter!
I cannot imagine difficult it would be to run S/360
era cables in
overhead trays. Many are an inch or more in diameter.
I've run lots of (heavy) 6/4 SOOW power cord on overhead cable trays, and
it is a pain, but it's way easier than pulling up half of the tiles in a
data center to figure out where the cable you're trying to pull is tangled,
just to find it stuck barely out of reach under the middle of some
production equipment.
An acceptable compromise is something like the 6 ft raised floor they have
over at NCSA. But the typical 12" raised floors I've had to deal with
are... unpleasant.
Pat
10:42 a.m.
The model 20 installations that I played with were mostly to supplant
unit-record gear, which typically did not use a raised floor
configuration. Mostly the CPU, card mulcher, printer and perhaps a 2311.
The installation auctioned off is one of the larger Model 20 setups that
I've seen.
--Chuck
1:30 p.m.
On 20/05/2019 18:42, Chuck Guzis via cctalk wrote:
The model 20 installations that I played with were
mostly to supplant
unit-record gear, which typically did not use a raised floor
configuration. Mostly the CPU, card mulcher, printer and perhaps a 2311.
The installation auctioned off is one of the larger Model 20 setups that
I've seen.
--Chuck
As Adam isn't on the list, and I am one of those in that photograph...
There was one 360/20 (pink) that was cabled up with 2 x 2311 disk, 2 x
2415 tape, 2560 MFCM and (I think) the 1403 printer. That system appears
to have been in use at that site. The second 360/20 (blue) was not
cabled to anything, and the 370/125 (yellow) likewise. So I expect
neither of those was used at all.
It can't have been a serious commercial installation, but maybe it was a
keen hobbyist who acquired the systems and decided to keep them running,
or perhaps used as a training tool. It apparently hasn't been used for
30+ years.
The "raised floor" was in a sorry state, the heavy units had sunk into
the soft wood panels. I guess the framing might have been useful.
It was good that all the punch cards required seem to be there and in
good condition: the Control Program (microcode) cards for the /20, RPG,
BAL etc. These need to be read - I know there are services that will
process scans, but is there anywhere (UK/Europe) that can take the
physical cards and give us back files?
The guys did a wonderful job, especially on that last day when they
managed to move most of the units and clear everything out. As
mentioned, the /20 CPUs and 2415 master unit were 500-600kg each and
about the limit of the tail-lift.
If anything I've written here is contradicted by anything in the blog,
believe the blog! https://ibms360.co.uk
--
Lawrence Wilkinson lawrence at ljw.me.uk
The IBM 360/30 page http://www.ljw.me.uk/ibm360
5:54 p.m.
On 5/20/19 1:30 PM, Lawrence Wilkinson via cctalk wrote:
As Adam isn't on the list, and I am one of those
in that photograph...
From the pictures, it looks like the packs were left in
the drives. You should pull them
and lock down the actuator arm before moving them
any further.
21 May
21 May
1:04 a.m.
First, congratulations from us :-)
Reading your blog, I'm glad that we did not win the auction. We simply do
not have the time and space (and money) to handle this, especially the
effort needed to get everything out of the house. I and Klemens both have
jobs that require us to do other things besides the museum ;-)
On Mon, 20 May 2019, Lawrence Wilkinson wrote:
There was one 360/20 (pink) that was cabled up with 2
x 2311 disk, 2 x 2415
tape, 2560 MFCM and (I think) the 1403 printer. That system appears to have
been in use at that site. The second 360/20 (blue) was not cabled to
anything, and the 370/125 (yellow) likewise. So I expect neither of those was
used at all.
Am I right that the 370/125 is missing something? At least there is
one yellow unit that is lacking the top half (or whatever).
These need to be read - I know there are services that
will process scans,
but is there anywhere (UK/Europe) that can take the physical cards and give
us back files?
Yes, we can (located in Stuttgart, not too far from Nuremburg). We've done
that with all cards that we got. Our setup uses the 1442 card reader on
the 1130 and a program that reads the cards and sends them via V.24 (with
a home-made tape punch interface to V.24 converter) to be captured by a
UNIX machine.
The guys did a wonderful job, especially on that last
day when they managed
to move most of the units and clear everything out. As mentioned, the /20
CPUs and 2415 master unit were 500-600kg each and about the limit of the
tail-lift.
I'm curious, did everything fit in a normal 7.5t truck or how do you move
all the stuff to the UK?
Christian
20 May
20 May
2:03 p.m.
On 5/20/19 10:40 AM, Patrick Finnegan via cctalk wrote:
An acceptable compromise is something like the 6 ft
raised floor they
have over at NCSA.
I've often wondered why more people don't do raised floors high enough
to walk under.
I guess they are trying to put it in one room and not have two separated
room stacked.
There's also the fact that having an open tile on a 6' (or taller)
raised floor is considerably more dangerous than something much shorter.
But the typical 12" raised floors I've had to
deal with are... unpleasant.
I'm used to 2' (24") raised floors.
--
Grant. . . .
unix || die
2:04 p.m.
try crawling under them
On Mon, May 20, 2019 at 4:03 PM Grant Taylor via cctalk <
cctalk at classiccmp.org> wrote:
On 5/20/19 10:40 AM, Patrick Finnegan via cctalk
wrote:
An acceptable compromise is something like the 6
ft raised floor they
have over at NCSA.
I've often wondered why more people don't do raised floors high enough
to walk under.
I guess they are trying to put it in one room and not have two separated
room stacked.
There's also the fact that having an open tile on a 6' (or taller)
raised floor is considerably more dangerous than something much shorter.
But the typical 12" raised floors I've
had to deal with are...
unpleasant.
I'm used to 2' (24") raised floors.
--
Grant. . . .
unix || die
2:25 p.m.
On 5/20/19 2:04 PM, Adrian Stoness via cctalk wrote:
try crawling under them
You'd only upset the mice and the cockroaches. I recall that back in
the day at CDC SVLOPS, the local CEs made a pet of one of the sub-floor
mice. Field crickets were a problem back then too--the moment that the
weather turned cool, they'd invade the facility. It was not unusual to
find a couple hiding out in one's desk drawers.
--Chuck
5:13 p.m.
had to crawl 300feet under raised floor cause it was carpeted.... runing
48pair fiber line took a good hour and half to get it over to the room with
the rack in it from the building raise
On Mon, May 20, 2019 at 4:25 PM Chuck Guzis via cctalk <
cctalk at classiccmp.org> wrote:
On 5/20/19 2:04 PM, Adrian Stoness via cctalk wrote:
try crawling under them
You'd only upset the mice and the cockroaches. I recall that back in
the day at CDC SVLOPS, the local CEs made a pet of one of the sub-floor
mice. Field crickets were a problem back then too--the moment that the
weather turned cool, they'd invade the facility. It was not unusual to
find a couple hiding out in one's desk drawers.
--Chuck
21 May
21 May
7:51 a.m.
Adrian Stoness wrote:
had to crawl 300feet under raised floor cause it was
carpeted.... runing
48pair fiber line took a good hour and half to get it over to the room with
the rack in it from the building raise
When I interned with the Social Security Administration, I was tasked with
crawling under the raised floor, pulling cable. As well as the various scenery
already mentioned by others, I came up at the other end with several dozen
9-track tape write rings. I stuck them on my (then skinny) arms as I crawled,
The older guys, waiting at the other end, laughed and told me stories of the
write ring battles that went on "back in the day".
Bill S.
---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus
2:21 p.m.
On Mon, May 20, 2019 at 5:07 PM William Donzelli via cctalk <
cctalk at classiccmp.org> wrote:
This gets at exactly the reason I dislike raised floors.... "out of sight -
out of mind". If all of the cabling is overhead, it gets ugly when you
abandon stuff, and there's more motivation to not make it a mess.
Also, I'd rather see all of the parts, rather than have them hidden from
view. I suppose that not everyone shares that point of view.
Pat
I'm used
to 2' (24") raised floors.
I was used to 24 inch floors packed with so many dead cables they were
effectively 6 to 12 inch floors. Or worse.
2:29 p.m.
On May 20, 2019, at 5:21 PM, Patrick Finnegan via
cctalk <cctalk at classiccmp.org> wrote:
On Mon, May 20, 2019 at 5:07 PM William Donzelli via cctalk <
cctalk at classiccmp.org> wrote:
This gets at exactly the reason I dislike raised floors.... "out of sight -
out of mind". If all of the cabling is overhead, it gets ugly when you
abandon stuff, and there's more motivation to not make it a mess.
Perhaps, but I remember seeing overhead cable trays running down the hallways in the DEC
Mill, full of ancient and obviously abandoned stuff.
I'm
used to 2' (24") raised floors.
I was used to 24 inch floors packed with so many dead cables they were
effectively 6 to 12 inch floors. Or worse.
Also, I'd rather see all of the parts, rather than
have them hidden from
view. I suppose that not everyone shares that point of view.
Raised floors often doubled as air handling space, supplying cold air for cooling the
equipment.
Overhead trays work well for rack mounted equipment, where the trays hang just above the
racks. When the equipment is in cabinets of varying height, like line printers or RP06
disk drives or stuff like that, running the cable way up in the air isn't so handy.
Never mind the fact that the connectors are configured to send the cable downward to the
floor, not up to the ceiling.
paul
9:53 a.m.
anyone figured out what these were being used for in that building?
On Mon, May 20, 2019 at 11:11 AM William Donzelli via cctalk <
cctalk at classiccmp.org> wrote:
I guess it
would look right for the era, but I'd never build a data
center
with raised flooring after my experiences with
them. It's such a pain to
work with compared to a sealed concrete floor and overhead cable trays.
But with a raised floor, you can whack the tile puller down in such a
way that it makes an enormously loud pop that startles everyone in the
datacenter!
I cannot imagine difficult it would be to run S/360 era cables in
overhead trays. Many are an inch or more in diameter.
--
Will
--
Will
21 May
21 May
3:26 a.m.
On Mon, 20 May 2019 at 23:10, Adrian Stoness via cctalk
<cctalk at classiccmp.org> wrote:
anyone figured out what these were being used for in that building?
Puma, the sportswear company, I think. Related to Adidas -- I believe
the companies were run by 2 brothers who fell out.
Puma was founded in Nuremberg. You can see Puma's logo in some of the
photos and it was mentioned in the eBay ad.
--
Liam Proven - Profile: https://about.me/liamproven
Email: lproven at cix.co.uk - Google Mail/Hangouts/Plus: lproven at gmail.com
Twitter/Facebook/Flickr: lproven - Skype/LinkedIn: liamproven
UK: +44 7939-087884 - ?R (+ WhatsApp/Telegram/Signal): +420 702 829 053
6:18 a.m.
The logo is sort of wrong for the era. I think it is just a sticker
that was applied.
And I do not think Puma would have been running on just a model 20,
and especially in such an "interesting" datacenter. By the 60s, they
were already a good sized company.
--
Will
On Tue, May 21, 2019 at 6:26 AM Liam Proven via cctalk
<cctalk at classiccmp.org> wrote:
On Mon, 20 May 2019 at 23:10, Adrian Stoness via cctalk
<cctalk at classiccmp.org> wrote:
anyone figured out what these were being used for in that building?
Puma, the sportswear company, I think. Related to Adidas -- I believe
the companies were run by 2 brothers who fell out.
Puma was founded in Nuremberg. You can see Puma's logo in some of the
photos and it was mentioned in the eBay ad.
--
Liam Proven - Profile: https://about.me/liamproven
Email: lproven at cix.co.uk - Google Mail/Hangouts/Plus: lproven at gmail.com
Twitter/Facebook/Flickr: lproven - Skype/LinkedIn: liamproven
UK: +44 7939-087884 - ?R (+ WhatsApp/Telegram/Signal): +420 702 829 053
20 May
20 May
2 p.m.
On 5/20/19 10:10 AM, William Donzelli via cctalk wrote:
I cannot imagine (how) difficult it would be to run
S/360 era cables in
overhead trays. Many are an inch or more in diameter.
I think it would be quite annoying to get cables from floor level up to
the overhead cable trays. Especially from all the equipment designed
for / assuming that cables went down below a raised floor.
Not the least of which would be burning (I'm guessing at least) 6' of
cable on each end.
--
Grant. . . .
unix || die
2:25 p.m.
-----Original Message-----
From: cctalk <cctalk-bounces at classiccmp.org> On Behalf Of Grant Taylor via
cctalk
Sent: 20 May 2019 22:00
To: cctalk at classiccmp.org
Subject: Re: Pleas ID this IBM system....
On 5/20/19 10:10 AM, William Donzelli via cctalk wrote:
IBM kit is designed to fit on raised floors. There isn't much clearance between the
bottom of the cabinets and the floor.
As someone has said the IBM bus and tag cables are thick and actual consist of multiple
co-ax cables in big bundle .
I have run a 4361 without raised floors, but we just ran the cables round the floor.
To run in them in roof level cable trays would be challenging. You would have to run them
across the floor for some distance before
Curving them up would be challenging as the minimum bend radius is large.
Dave
I cannot imagine (how) difficult it would be to
run S/360 era cables
in overhead trays. Many are an inch or more in diameter.
I think it would be quite annoying to get cables from floor level up to the
overhead cable trays. Especially from all the equipment designed for /
assuming that cables went down below a raised floor.
Not the least of which would be burning (I'm guessing at least) 6' of cable on
each end.
--
Grant. . . .
unix || die
21 May
21 May
1:13 a.m.
On Mon, 20 May 2019, Patrick Finnegan wrote:
I guess it would look right for the era, but I'd
never build a data center
with raised flooring after my experiences with them. It's such a pain to
work with compared to a sealed concrete floor and overhead cable trays.
My experience is that overhead trays are not better. We have rooms with
only raised floors, and one with both. You definitely need a raised floor
for a data center. You need it for forced air cooling and for running the
water and condensate pipes. And overhead trays are much more difficult to
work with if you want to lay new cables because you have to climb up and
down the ladder all the time, moving the ladder from here to there and
back to here...
Christian
22 May
22 May
2:33 a.m.
On Wed, 22 May 2019 at 05:17, Mark Linimon via cctalk
<cctalk at classiccmp.org> wrote:
On Tue, May 21, 2019 at 09:52:19AM -0600, Grant Taylor via cctech wrote:
FWIW, I had no clue either. I was hoping to work it out from context.
No joy so far.
--
Liam Proven - Profile: https://about.me/liamproven
Email: lproven at cix.co.uk - Google Mail/Hangouts/Plus: lproven at gmail.com
Twitter/Facebook/Flickr: lproven - Skype/LinkedIn: liamproven
UK: +44 7939-087884 - ?R (+ WhatsApp/Telegram/Signal): +420 702 829 053
I think Google and their YAWNs
Definition, please? Wikipedia and Urban Dictionary are no help. A Google
search itself is nothing but false positives.
12:38 a.m.
On Tue, 21 May 2019, Grant Taylor wrote:
On 5/21/19 2:13 AM, Christian Corti via cctalk wrote:
I'm not Google, I can't put my infrastructure in the Arctic to reduce my
cooling needs.
Christian
You definitely need a raised floor for a data
center.
I think Google and their YAWNs will disagree with you on an actual /need/
for
a raised floor in a data center.
8:45 a.m.
On 5/22/19 1:38 AM, Christian Corti via cctalk wrote:
I'm not Google, I can't put my infrastructure
in the Arctic to reduce my
cooling needs.
Google has infrastructure all over the world. Most of the locations
require active cooling year round. A quick Google search turns up the
following maps:
Link - Map of all Google data center locations
- https://royal.pingdom.com/map-of-all-google-data-center-locations/
There may be a few data centers in the northern hemisphere that are in a
temperate ~ cold climate. But I don't see any in the Arctic.
Also, Google wants the infrastructure as close to the end users as
reasonably possible.
--
Grant. . . .
unix || die
21 May
21 May
9:38 a.m.
On Tue, May 21, 2019, 04:13 Christian Corti via cctalk <
cctalk at classiccmp.org> wrote:
You definitely need a raised floor
for a data center. You need it for forced air cooling and for running the
water and condensate pipes.
Ductwork doesn't have to be below the floor. Modern co-lo facilities that I
have been in (such as Switch Supernap) don't have a raised floor.
Plumbing (unless you're doing aisle containment or RDHx) shouldn't run
through the IT space in the data center.
Cooling water to racks should be dewpoint adjusted, so you don't need
condensate drains inside the DC.
And overhead trays are much more difficult to
work with if you want to lay new cables because you
have to climb up and
down the ladder all the time, moving the ladder from here to there and
back to here...
I solved that by having multiple ladders. In my experience, it's a lot
easier than trying to reach through a cluttered raised floor under racks.
The only good reason that I have seen in this thread for a raised floor is
to match older equipment that routes cables downwards.
Pat
10:03 a.m.
all systems have their advantages disadvantages it all depends on what your
doing and designs u choose. personaly i think raised floor and tray above
are best then u keep all ur power below away from ur data lines plus but
then ur setup is only as good as the lazyest tech u get comming in running
stuff.
On Tue, May 21, 2019 at 11:38 AM Patrick Finnegan via cctalk <
cctalk at classiccmp.org> wrote:
On Tue, May 21, 2019, 04:13 Christian Corti via cctalk
<
cctalk at classiccmp.org> wrote:
You definitely need a raised floor
for a data center. You need it for forced air cooling and for running the
water and condensate pipes.
Ductwork doesn't have to be below the floor. Modern co-lo facilities that I
have been in (such as Switch Supernap) don't have a raised floor.
Plumbing (unless you're doing aisle containment or RDHx) shouldn't run
through the IT space in the data center.
Cooling water to racks should be dewpoint adjusted, so you don't need
condensate drains inside the DC.
And overhead trays are much more difficult to
work with if you want to lay new cables because
you have to climb up and
down the ladder all the time, moving the ladder from here to there and
back to here...
I solved that by having multiple ladders. In my experience, it's a lot
easier than trying to reach through a cluttered raised floor under racks.
The only good reason that I have seen in this thread for a raised floor is
to match older equipment that routes cables downwards.
Pat
10:43 a.m.
No modern datacenter that I have seen still uses a raised floor *OTHER THAN* about 3
inches for a ground plane. There is a reason for that... the old idea of forced cooling
under the floor and mixing power & data cables there has been found to be a truly bad
idea.
Power in most any modern datacenter is via "track lighting" rails directly over
the row of racks. Ex: https://www.starlinepower.com/busway/
HVAC units are generally not inside the datacenter any longer either, nor are UPS's.
These cause access control issues with vendors doing maintenance and raises issues for
publicly traded companies that must meet certain control & accountability standards.
Not to mention locating them on the datacenter floor can cause issues with cool air
routing. So these days the hot and cold aisles are alternating, with a windsock/tube above
the racks for cold air delivery to the front and an 8 foot or so fan built in the hot
aisle wall. This is the optimum place for particle (smoke) sensors as well. That design
also lets in certain situations outside air to be used.
So to say you "definitely need a raised floor for a datacenter" is only true if
you are trying to create a period replica. A modern datacenter built that way... well...
not sure any self-respecting contractor would build one 'the old way' :)
J
-----Original Message-----
From: cctalk [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Adrian Stoness via
cctalk
Sent: Tuesday, May 21, 2019 12:03 PM
To: Patrick Finnegan <pat at vax11.net>; General Discussion: On-Topic and Off-Topic
Posts <cctalk at classiccmp.org>
Subject: Re: Pleas ID this IBM system....
all systems have their advantages disadvantages it all depends on what your doing and
designs u choose. personaly i think raised floor and tray above are best then u keep all
ur power below away from ur data lines plus but then ur setup is only as good as the
lazyest tech u get comming in running stuff.
On Tue, May 21, 2019 at 11:38 AM Patrick Finnegan via cctalk < cctalk at
classiccmp.org> wrote:
On Tue, May 21, 2019, 04:13 Christian Corti via cctalk
<
cctalk at classiccmp.org> wrote:
You definitely need a raised floor
for a data center. You need it for forced air cooling and for
running the water and condensate pipes.
Ductwork doesn't have to be below the floor. Modern co-lo facilities
that I have been in (such as Switch Supernap) don't have a raised floor.
Plumbing (unless you're doing aisle containment or RDHx) shouldn't run
through the IT space in the data center.
Cooling water to racks should be dewpoint adjusted, so you don't need
condensate drains inside the DC.
And overhead trays are much more difficult to
work with if you want to lay new cables because
you have to climb up
and down the ladder all the time, moving the ladder from here to
there and back to here...
I solved that by having multiple ladders. In my experience, it's a lot
easier than trying to reach through a cluttered raised floor under racks.
The only good reason that I have seen in this thread for a raised
floor is to match older equipment that routes cables downwards.
Pat
11:04 a.m.
It was thus said that the Great Jay West via cctalk once stated:
No modern datacenter that I have seen still uses a
raised floor *OTHER
THAN* about 3 inches for a ground plane. There is a reason for that... the
old idea of forced cooling under the floor and mixing power & data cables
there has been found to be a truly bad idea.
I recall walking through the former IBM main site in Boca Raton [1] in the
late 90s with a few friends. We were amazed to see labs with six-foot deep
pits---the tiles long gone and all cables removed. It was quite the sight
to see.
-spc
[1] Google street view:
https://www.google.com/maps/@26.3912323,-80.1059743,3a,75y,272.91h,88.2t/da…
The building is rated for a class-5 hurricane. It's quite
impressive.
12:34 p.m.
I might think far more "obsolete" than "bad idea". It worked very
well
for the mainframe folks.
--
Will
On Tue, May 21, 2019 at 1:50 PM Jay West via cctalk
<cctalk at classiccmp.org> wrote:
No modern datacenter that I have seen still uses a raised floor *OTHER THAN* about 3
inches for a ground plane. There is a reason for that... the old idea of forced cooling
under the floor and mixing power & data cables there has been found to be a truly bad
idea.
Power in most any modern datacenter is via "track lighting" rails directly over
the row of racks. Ex: https://www.starlinepower.com/busway/
HVAC units are generally not inside the datacenter any longer either, nor are UPS's.
These cause access control issues with vendors doing maintenance and raises issues for
publicly traded companies that must meet certain control & accountability standards.
Not to mention locating them on the datacenter floor can cause issues with cool air
routing. So these days the hot and cold aisles are alternating, with a windsock/tube above
the racks for cold air delivery to the front and an 8 foot or so fan built in the hot
aisle wall. This is the optimum place for particle (smoke) sensors as well. That design
also lets in certain situations outside air to be used.
So to say you "definitely need a raised floor for a datacenter" is only true if
you are trying to create a period replica. A modern datacenter built that way... well...
not sure any self-respecting contractor would build one 'the old way' :)
J
-----Original Message-----
From: cctalk [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Adrian Stoness via
cctalk
Sent: Tuesday, May 21, 2019 12:03 PM
To: Patrick Finnegan <pat at vax11.net>; General Discussion: On-Topic and Off-Topic
Posts <cctalk at classiccmp.org>
Subject: Re: Pleas ID this IBM system....
all systems have their advantages disadvantages it all depends on what your doing and
designs u choose. personaly i think raised floor and tray above are best then u keep all
ur power below away from ur data lines plus but then ur setup is only as good as the
lazyest tech u get comming in running stuff.
On Tue, May 21, 2019 at 11:38 AM Patrick Finnegan via cctalk < cctalk at
classiccmp.org> wrote:
On Tue, May 21, 2019, 04:13 Christian Corti via
cctalk <
cctalk at classiccmp.org> wrote:
You definitely need a raised floor
for a data center. You need it for forced air cooling and for
running the water and condensate pipes.
Ductwork doesn't have to be below the floor. Modern co-lo facilities
that I have been in (such as Switch Supernap) don't have a raised floor.
Plumbing (unless you're doing aisle containment or RDHx) shouldn't run
through the IT space in the data center.
Cooling water to racks should be dewpoint adjusted, so you don't need
condensate drains inside the DC.
And overhead trays are much more difficult to
work with if you want to lay new cables because
you have to climb up
and down the ladder all the time, moving the ladder from here to
there and back to here...
I solved that by having multiple ladders. In my experience, it's a lot
easier than trying to reach through a cluttered raised floor under racks.
The only good reason that I have seen in this thread for a raised
floor is to match older equipment that routes cables downwards.
Pat
12:51 p.m.
On 5/21/19 12:34 PM, William Donzelli via cctalk wrote:
I might think far more "obsolete" than
"bad idea". It worked very well
for the mainframe folks.
...except when it didn't. On more than one occasion, I recall watching
some poor soul with a cart balanced with long (3 foot) trays full of
cards, hitting a lifted separator strip in the raised floor. Over goes
the cart, the cards spread in every direction, and the pusher sitting in
the middle of it all, close to tears.
If you happen to chance across any old card decks and note traces of a
diagonal line drawn across the top, that's the reason--a visual aid for
reassembling a spilled deck.
I think the integration folks eventually moved to 4-wheel carts instead
of the 3-wheel ones for exactly this reason. Their job was integrating
new code and patches into the existing product base. The trip with the
cards was to make a tape of the data, so that the trip didn't have to be
repeated.
--Chuck
12:55 p.m.
You can't blame anyone but they idiot using a 3 wheel cart while
moving decks of cards. No sympathy from me.
--
Will
On Tue, May 21, 2019 at 3:51 PM Chuck Guzis via cctalk
<cctalk at classiccmp.org> wrote:
On 5/21/19 12:34 PM, William Donzelli via cctalk wrote:
I might think far more "obsolete" than
"bad idea". It worked very well
for the mainframe folks.
...except when it didn't. On more than one occasion, I recall watching
some poor soul with a cart balanced with long (3 foot) trays full of
cards, hitting a lifted separator strip in the raised floor. Over goes
the cart, the cards spread in every direction, and the pusher sitting in
the middle of it all, close to tears.
If you happen to chance across any old card decks and note traces of a
diagonal line drawn across the top, that's the reason--a visual aid for
reassembling a spilled deck.
I think the integration folks eventually moved to 4-wheel carts instead
of the 3-wheel ones for exactly this reason. Their job was integrating
new code and patches into the existing product base. The trip with the
cards was to make a tape of the data, so that the trip didn't have to be
repeated.
--Chuck
1:11 p.m.
On 5/21/19 12:55 PM, William Donzelli wrote:
You can't blame anyone but they idiot using a 3
wheel cart while
moving decks of cards. No sympathy from me.
They were using what was available. I&R bought their own four-wheel cart
(looked like a standard gray industrial shop cart that you can purchase
today. I suspect the small ones were intended for easy access when
delivering cards, tapes and listings to many people and the easy ability
to get in and out of elevators and through doorways. I suspect that they
were repurposed small mail carts.
--Chuck
1:17 p.m.
On 5/21/19 1:51 PM, Chuck Guzis via cctalk wrote:
...except when it didn't. On more than one
occasion, I recall watching
some poor soul with a cart balanced with long (3 foot) trays full of
cards, hitting a lifted separator strip in the raised floor. Over goes
the cart, the cards spread in every direction, and the pusher sitting
in the middle of it all, close to tears.
I'm sure that was /a/ problem. But I'm not comfortable attributing that
problem to the raised floor.
I expect that the same problem would be effected by an elevator that
doesn't stop perfectly level with the floor, or has too wide a gap
between the car and the floor, or even on tiled floor.
That really seems to me like it's a sub-optimal design, pushed past it's
operating parameters by overloading it.
I can't fault the raised floor for that problem.
--
Grant. . . .
unix || die
2:33 p.m.
On 5/21/19 1:17 PM, Grant Taylor via cctalk wrote:
I'm sure that was /a/ problem.? But I'm not
comfortable attributing that
problem to the raised floor.
I expect that the same problem would be effected by an elevator that
doesn't stop perfectly level with the floor, or has too wide a gap
between the car and the floor, or even on tiled floor.
That really seems to me like it's a sub-optimal design, pushed past it's
operating parameters by overloading it.
I can't fault the raised floor for that problem.
No, but it's just one of the anecdotes that go with the terrain, like
leaving a box or two of cards or a stack of tapes or your dinner on top
of a 1403N1 when it ran out of forms..
How well sealed were the raised floors? I ask this because i recall an
episode or two where a disk drive would spring a leak and make a
beautiful slippery pool on the floor, just waiting for the next operator
to dash by. I always wondered how much of the stuff made it to the subfloor.
--Chuck
2:37 p.m.
On 5/21/19 3:33 PM, Chuck Guzis via cctalk wrote:
How well sealed were the raised floors? I ask this
because i recall an
episode or two where a disk drive would spring a leak and make a beautiful
slippery pool on the floor, just waiting for the next operator to dash
by. I always wondered how much of the stuff made it to the subfloor.
I wouldn't consider them sealed at all.
Is suspect that vast majority of the fluid would spread out across the
tiles than would actually drip through. I expect that it would have a
lot more to do with the rate of release to determine if it would flow
across tile gaps vs drip down in between.
--
Grant. . . .
unix || die
3:03 p.m.
Chuck Guzis wrote:
How well sealed were the raised floors?
If you mean "when properly installed", pretty well.
If you mean "as actually used", not at all.
With the exception of the very small raised floor under
the HP-3000 in my high school, every place I've encountered
raised floors there were "permanently removed" tiles, tiles
with corners cut off, tiles with large holes drilled into their
centers, etc. etc.
Bill S.
---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus
5:32 p.m.
On 05/21/2019 04:33 PM, Chuck Guzis via cctalk wrote:
How well sealed were the raised floors?
Not at
all. They were 2 foot squares sitting on adjustable
pillars. Each pillar supported 4 tiles where they all met
at the corners. You could easily slip a punch card (or
credit card) between most of the tiles. Anything poured on
the floor would run down between the tiles.
Jon
11:22 a.m.
On 5/21/2019 11:03 AM, Adrian Stoness via cctalk wrote:
all systems have their advantages disadvantages it all
depends on what your
doing and designs u choose. personaly i think raised floor and tray above
are best then u keep all ur power below away from ur data lines plus but
then ur setup is only as good as the lazyest tech u get comming in running
stuff.
I say use the STAR TREK layout. All cables run under the AIR DUCTS, and
you crawl through those. :)
Ben.
10:28 a.m.
"1 in the elvator takes you to the data center.
B in the elvator takes you to the tall enough to walk in raised floor.
M or 2 in the elevator takes you to the overhead ca ble area."
22 May
22 May
12:49 a.m.
On Tue, 21 May 2019, Patrick Finnegan wrote:
Plumbing (unless you're doing aisle containment or
RDHx) shouldn't run
through the IT space in the data center.
So how exactly do you attach a modern water cooled rack system to your
cooling water system if not using plumbing?
Cooling water to racks should be dewpoint adjusted, so
you don't need
condensate drains inside the DC.
That is a requirement, you must have condensate drains just in case that
the room temperature and/or air humidity rises, or cooling water
temperature goes below the dew point.
Christian
1:24 a.m.
On 5/22/19 12:49 AM, Christian Corti via cctalk wrote:
On Tue, 21 May 2019, Patrick Finnegan wrote:
So how are data centers cooled with water now? Does the water cool
coldplates directly?
I've had only a couple of instances where cooling water was used. In
the case of CDC mainframes, it was used to cool the condenser coils in
the refrigeration units (located in the mainframe). I believe that Cray
initially used the same guy that CDC used to fabricate the cooling tubing.
I recall visiting the Honeywell plant in Phoenix not long after they
took it over from GE and the engineers there were tinkering with a
direct water-cooling setup--water circulated in each rack (connected by
what was probably vinyl tubing, I don't recall, only that it was
translucent), with copper diaphragms serving as the interface between
the water and the semiconductors. I recall from comments made that
algae was a problem and adding an algicide to the cooling water tended
to corrode the copper diaphragms.
To the best of my knowledge, this was a test setup--it certainly had an
impressive instrumentation unit with multiple CRTs and neon
thermometer-type displays.
The most extreme example I ever ran into of cooling was the ETA-10
supercomputer--cooled with liquid nitrogen, supplied by a cryostat. I
don't recall what cooled the latter.
--Chuck
Plumbing (unless you're doing aisle
containment or RDHx) shouldn't run
through the IT space in the data center.
So how exactly do you attach a modern water cooled rack system to your
cooling water system if not using plumbing?
8:30 a.m.
On Wed, May 22, 2019 at 4:24 AM Chuck Guzis via cctalk <
cctalk at classiccmp.org> wrote:
On 5/22/19 12:49 AM, Christian Corti via cctalk
wrote:
That's an option. I support 20-30kW/rack systems with using Coolcentric
(passive) rear door heat exchangers, which have a dewpoint-adjusted cooling
loop. The air is generally managed using CRAC units / building air
handlers.
On Tue, 21 May 2019, Patrick Finnegan wrote:
So how are data centers cooled with water now? Does the water cool
coldplates directly?
Plumbing (unless you're doing aisle
containment or RDHx) shouldn't run
through the IT space in the data center.
So how exactly do you attach a modern water cooled rack system to your
cooling water system if not using plumbing? I recall visiting the Honeywell plant in Phoenix not
long after they
took it over from GE and the engineers there were tinkering with a
direct water-cooling setup--water circulated in each rack (connected by
what was probably vinyl tubing, I don't recall, only that it was
translucent), with copper diaphragms serving as the interface between
the water and the semiconductors. I recall from comments made that
algae was a problem and adding an algicide to the cooling water tended
to corrode the copper diaphragms.
New versions of that are made by companies such as Cool-IT, or HPE's SGI
systems. The materials used have progressed quite a bit, mostly
eliminating the algae and corrosion problems, and people have mostly
settled on ~25-40C (77-104F) water for cooling, to avoid condensing
humidity.
Pat
8:30 a.m.
New subject: Raised Floors
On Wed, 22 May 2019, Chuck Guzis wrote:
So how are data centers cooled with water now? Does
the water cool
coldplates directly?
Sort of, see for example the Rittal LCP system.
I've had only a couple of instances where cooling
water was used. In
All current high-density rack systems need liquid cooling. I want to
remind that a single rack may produce 10kW of heat or more.
the case of CDC mainframes, it was used to cool the
condenser coils in
the refrigeration units (located in the mainframe). I believe that Cray
initially used the same guy that CDC used to fabricate the cooling tubing.
That is how todays rack systems work.
Christian
8:23 a.m.
On Wed, May 22, 2019 at 3:50 AM Christian Corti via cctalk <
cctalk at classiccmp.org> wrote:
For what we do, It's not, we don't do that, and our equipment suppliers
don't suggest it. The air is humidity controlled before it gets near the
IT racks, by building air handlers and CRAC units.
Our process water loop inside the room watches the humidity and adjusts the
water temperature appropriately. It'd be hard to for the dewpoint to
change rapidly enough to cause condensation vs how quickly systems can
increase the water temperature to compensate.
Modern systems using direct-to-system water cooling also tend to use "warm
water" cooling, since they don't need water below ~90F (~32C) to cool
effectively (some specify up to ~104F (~40C). RIP the data center that has
a dewpoint that high, since the servers will probably be covered in rust
pretty quickly.
Pat
On Tue, 21 May 2019, Patrick Finnegan wrote:
I guess I should have added "etc" to my exception list? My point was that
air handlers, CRACs, etc, should not have their plumbing running through
the data center since they should be outside in a mechanical room anyways.
Plumbing (unless you're doing aisle
containment or RDHx) shouldn't run
through the IT space in the data center.
So how exactly do you attach a modern water cooled rack system to your
cooling water system if not using plumbing?
Cooling water
to racks should be dewpoint adjusted, so you don't need
condensate drains inside the DC.
That is a requirement, you must have condensate drains just in case that
the room temperature and/or air humidity rises, or cooling water
temperature goes below the dew point.
1822
days inactive
1824
days old
46 comments
19 participants
participants (19)
-
aek＠bitsavers.org -
bfranchuk＠jetnet.ab.ca -
cc＠informatik.uni-stuttgart.de -
cclist＠sydex.com -
cctalk＠gtaylor.tnetconsulting.net -
cisin＠xenosoft.com -
dave.g4ugm＠gmail.com -
elson＠pico-systems.com -
jnc＠mercury.lcs.mit.edu -
jwest＠classiccmp.org -
linimon＠lonesome.com -
ljw-cctech＠ljw.me.uk -
lproven＠gmail.com -
pat＠vax11.net -
paulkoning＠comcast.net -
spc＠conman.org -
tdk.knight＠gmail.com -
wdonzelli＠gmail.com -
wh.sudbrink＠verizon.net