Getting rid of the powder printer and blade servers

Space and infrastructure
1

As of the latest QGM today, we have set a tentative move day of December 13th, depending on circumstances.

This is a little over one month from today, and I am proposing that we clear some necessary space.

The two elephants in the room (as I see them) are the District 9 server blades and the powder printer.

The printer

This has been in the space for a year and a half without being run once. There have been at least three people (that I know of) on point in trying to get it working, but they have all moved on to other things. The intention was that once it starts working, it will draw in members. After looking into it, I disagree, and I don’t believe it is an asset to VHS.
Reference: tool:z-corporation_z402_3d_printer [Vancouver Hack Space]

Cons:

  • Has never been used/fixed
  • Will require significant repairs and maintenance that are currently outside the wheelhouse of any of our members
  • Is huge and heavy
  • Requires expensive consumables
  • Does not print significantly better results than our (working) FDM printers (Source: 1 2)
  • Does not significantly draw in more members than our actual working printers

I do not see any pros in keeping this printer at this time.

Server Blades

I have never seen anyone working on these or show any interest in getting them useful. I’m not sure how they are useful.

Further Action

If there is nothing anyone would like to discuss, I will attempt to list these on Craigslist and donate any profits to the VHS moving fund. If I cannot get any interest in buyers, I will do what is required to e-waste them responsibly.

I anticipate that there will be some dissent, though, so I propose this:

We do not move either of these two pieces of equipment if they are not working by the time that we intend to do the bulk of the moving (December 13th). That is one month.

If there are no objections, I will move on this immediately.

2 Likes
2

I would like to harvest the 3D stage from the powder printer if we are going to trash it or don’t find someone to buy it from us.

Awesome! there are a few other items i would love to add to this list. I will take pictures and reply to this thread later on. Selling things that no one uses or we have much better versions makes perfect sense to me.

I agree with this proposal.

1 Like
3

I know it’s a pain in the butt to move, but i’d be interested in at least taking a look at the 3d powder printer in Feb if people would use it if I fixed it. I believe I have the required skills to fix it, if not the funds to do so. I agree that its not likely to attract many members though, its certainly got a very slow build speed.

4

Hey Mark,

At this point, we’re looking to have it either gone or working by December. We simply do not have the resources to lug around a 400lb paper-weight.

Sorry :frowning:

1 Like
5

The printer needs replacement printer heads

http://vanhack.ca/doku.php?id=tool:z-corporation_z402_3d_printer&s=3d&s=printer

.

Here is a a link on replacing the printer headshttp://blog.freesideatlanta.org/2011/07/making-print-cartridges-for-z400z402-3d.html

Canon BC-20 cartridges (must be black to work)

Daniel"

Some more background about the printer not included in the wiki

Okay much over due update.

Finally had an hour to get it working. Yes getting the HD mirrored is

the first step. It was the one thing the original owner warned to do

before proceeding. I made 2 mirrors and then tried both of them

yesterday. Neither worked. Original how every still does. So the

mirroring process was the failure. I know have those hd’s back home

and will attempt to create new mirrors to get it working.

I made an exact mirror copy so that may actually be the problem so

I’ve 3 hd’s maybe a 4th I’ll try different versions then be in later

this week.

The rest of the machine runs fine. BUT god is it old.

Almost all of the parts are off the shelf accept for the controller

card which is ISA. So no upgrading the motherboard, etc, atm.

That power supply I will replace just because its only a 250 and

around 12 years old.

A lot of the tubes for the fluid look like they need replacing. We did

receive a donation of high grade medical hoses a month ago from Mark

(tks again)

High Quality stuff with some really good connectors so if anything an

improvement over whats there.

The Broken Printer head is the reason the original owner donated it to

us in the first place. A while back I found a way to replace it with a

bubble jet printer head. But I lost that page and have not found it

again.

"

6

I support getting rid of this printer. It’s not really a good fit for out space even if it was working. They are a bit messy, don’t make a very strong object (is suitable for making cast forms however) and require quite a bit of maintenance including modifying Canon BC-20 cartridges into Z Corp cartridges. Without a very enthusiastic champion, it’s really just a waste for us to have it.

7

Tim,

I have that email, too. It was sent in July, and now it is November. We know (approximately) what needs to be done to fix it, but I have to assume that no one is interested.

There are many useful projects at VHS for someone look for one. I don’t believe that this is useful, and it is one of the largest ongoing projects in the space, and the only one that has been stagnant almost the entire time we have been at The Bunker.

1 Like
8

my purpose was to inform people of the current status and en-power them.

2 Likes
9

I just want to chime in here and note that when we picked the powder printer up from its original owners, we saw some samples of its print quality - and from memory it was far superior to what I’ve seen from the FDM printers we have. The other consideration is that it can do much more complex (true) 3D shapes due to the inherent support of the raw material surrounding the object as it prints.

I won’t fight to keep it (as I’m unlikely to use it, and certainly can’t take on fixing/maintaining/supporting another large tool), but would be disappointed to see it go.

10

Did I miss something or have we found a new space?

Paul S. Gill

VE7PXG

604.250.0060

11

Into a new space or into storage, either way we need to move

12

If anyone has dibs on that powder printer first, its me. Having, yes I know not gotten it to work yet, but having put more hours then anyone else into it to try. I’ll find a place to store the thing and continue to work on it.

1 Like
13

Current State of the Blade Server Cluster(?)

I think the District 9 blades are missing their RAM. I took a look at one of the blades a few days ago. I think each blade accepts 4 slots of desktop DDR2 RAM, however I could be wrong, and it’s possible that it requires buffered DIMMs or something like that.

In order to do something useful, each blade will require about 512 MB of RAM at the minimum, otherwise there really isn’t much room for cluster computation applications to play around in without swapping to disk.

Someone seems to have designed (or at least tried to design) a custom power supply for the blades. I have no clue if it works or not. There are some pieces of paper on the rack indicating progress, and they say that the project to get them running was started in early 2013.

Actually using the cluster to do things:

After the hardware, then there’s the question of software, drivers, networking, and how to orchestrate tasks amongst the cluster. The way I see it, there are essentially two ways to use the cluster.

The first way is to use it the same way that supercomputers are used. I’m thinking that a cluster management workstation would be needed to actually issue setup commands from a central location, transmit work problems over 10 Gigabit Ethernet or Infiniband, and then receive all the answers in one place. There is some API which I can’t remember the name of at the moment which is used by a lot of universities for controlling cluster jobs, map-reduce problems, message passing stuff, etc. Basically you would need to design cluster jobs so that they operate on very small chunks of work at once, which are dispersed through the cluster, and the results are assembled into larger chunks of work, and so on, until you get the answer to your question (which is 42 of course!) This method works for things like fluid dynamics simulations, electromagnetic simulations, etc. The other way is to have your app function as a series of autonomous agents which send tiny pieces of information to each other randomly in order to form a distributed computation network. This is a method commonly used by Erlang and Scala/Akka programmers for creating extremely complex systems with Byzantine fault tolerance, hot code reloading, and live migration.

If the cluster is capable of being set up for virtualization, then that could be an alternative way of allocating and managing cluster computing resources. Something like VMWare ESXi, Xen, or SmartOS might work. However this limits the potential of the cluster since a single VM can’t really span across the entire cluster, you’d end up cutting the cluster into many small VMs, which would at best have live migration and/or live redundancy capabilities. A storage array would also be needed since each blade only has a single small 2.5" hard drive inside it, and that storage array would have to be attached via a high speed interface (10GbE, IB, FC, or SAS). VMs really love having fast storage.

The advantage of the VM method is that it is simple to understand and use, since you use it just like a normal computer, you can spin up new ones whenever you want, and they are isolated from each other. However it is less efficient at using the capabilities of the cluster.

So whatever happens to the blade servers, it seems that they will require a significant investment in time, space, extra parts, and several kilowatts of electricity to get them up to a usable state. (I’m guessing about 10 kilowatts max, which means that a 240 volt 40 amp stove plug should be able to power the whole thing in theory…)

14

I am okay with this

Thanks for the update. So in your opinion they are not worth getting going?

I expect @toma and possibly @SDY to weigh in here at some point.

edit: Farrell ( @toptekkie) was also looking at the powder printer last week, too, so he might want to comment

15

Daniel,

Do you have the Canon BC-20 cartridges ordered?If not I will order these and contribute them . How many?

16

4 blade chassis - I believe IBM Bladecenter E
14 blades per chassis

28 hs20 blades
2 CPUs - Intel Xeon 2.8ghz
4 gb of ram
40gb hdd

28 ls20 blades
2 CPUs - Amd opteron 2.2ghz DualCore
4 gb of ram
40gb hdd

1 Like
17

Getting rid of the D9 and the powder printer would make me so relieved.
I’ve been not-so-quietly against them from day 1.

18

No I have not yet gotten any replacement canon cartridges. There was an offer from someone to keep an eye out for them couple months ago… I have not heard anything since then. Used cartridges are better then full ones, As its only the electronics and the nozzles that the printer needs. Not the ink.

Farrell and I did have a discussion about the printer last week after my 3d modelling class.

19

If they have not been ordered, they are not.

How many do you want me to buy for you?

20

Have you gotten a chance to put these servers on Craigslist yet?
The E-Waste is getting picked up tomorrow. I don’t think we have time to include it with this pick up.