[Halld-online] Fwd: Reconfiguration of gluon machines

[David Lawrence]

Hi All,

  It looks like there is a small delay in getting the new gluon41-43 computers on the high-speed 
networks. Thus, they cannot be used for DAQ operations at the moment. Another announcement
will be sent out once they are available. Let me know if you a node or nodes temporarily assigned
to you.

Regards,
-David

Begin forwarded message:

> From: Paul Letta <letta at jlab.org>
> Subject: Re: Reconfiguration of gluon machines
> Date: September 29, 2014 at 3:51:04 PM EDT
> To: David Lawrence <davidl at jlab.org>, David Abbott <abbottd at jlab.org>, Hovanes Egiyan <hovanes at jlab.org>
> 
> I removed the dual port IB cards from 50,51,52 and did all the renaming below.  I also moved the chassis that holds 100-103 over to be with the other farm nodes.  
> 
> However, I was not able to put the IB cards into gluon41,42,43.   These are 1U boxes that require full height PCI cards.  The IB cards that came out of the supermicro blades are 1/2 height cards.  I just need some full height dual port IB face plates.. not new cards..  I will look into getting them ASAP.   So for now, gluon40 has all the IB ports, while gluon41,42,43 only have the 1 onboard IB port.
> 
> Dave Abbott -- I left the IB cables with the EMUs (gluon40+)... so there are no IB cables going to the newly named farm nodes gluon100,101,102,103.  You will need to stop over and do what ever IB cabling you want done.  (or let me know when you are coming and I will come to help..)
> 
> 
> 
> Paul
> 
> From: "Paul Letta" <letta at jlab.org>
> To: "David Lawrence" <davidl at jlab.org>
> Sent: Thursday, September 25, 2014 4:22:49 PM
> Subject: Re: Reconfiguration of gluon machines
> 
> Thanks for the the detailed plan.  This will not be too bad...  I'll need the outage time... probably 3-4 hours...  I am out Friday and Monday though...
> 
> Also -- instead of going to gluon113+ .. there is a gap at 100,101,102,103.... how about moving them there..
> 
> That would leave the farm nodes as:
> 100   102 (new farm chassis from EMUs)(50->100, 51->101, 52->102, 112->103)
> 101   103
> 
> 104  106   (current top farm chassis) (NO change)
> 105  107
> 
> 108  110   (current 2nd farm chassis)(NO change)
> 109  111
> 
> 
> And the EMUs nodes would be:
> 40  (AMD 1U's)
> 41
> 42
> 43
> 
> 46   48   (current emu chassis)(NO change)
> 47   49
>  
> 55  56  (2 borrowed chassis') (NO change)
> 57
> 
> 
> From: "David Lawrence" <davidl at jlab.org>
> To: "Paul Letta" <letta at jlab.org>
> Cc: "Sergey Furletov" <furletov at jlab.org>, "David Abbott" <abbottd at jlab.org>, "Alex Somov" <somov at jlab.org>, "Hovanes Egiyan" <hovanes at jlab.org>, "Eugene Chudakov" <gen at jlab.org>
> Sent: Thursday, September 25, 2014 3:59:13 PM
> Subject: Reconfiguration of gluon machines
> 
> Hi Paul,
> 
>   I spoke with both Dave A. and Sergey F. about configuring the new AMD machines with
> dual port Mellanox cards. I think this may be fairly close to what you and Dave discussed.
> 
> Here is the proposed plan:
> 
> 1.) move the dual port cards from gluon50-gluon52 into gluon41-gluon43
> 
> 2.) rename gluon50-gluon52 to be gluon113-gluon115
> 
> 3.) rename gluon53 to gluon40
> 
> 4.) move the gluon112-gluon115 chassis to the same rack as the other farm nodes
>     but *only* if it is not too much trouble. If the cables are too short or are difficult
>     to re-route, just leave it where it is
> 
> 5.) move the gluon112 blade into the correct position in its chassis (upper-left)
>     so all the same blades in the chassis will follow the same ordering as the others
> 
> This will leave us with no (new) gluon50 machines and 8 gluon40 machines (4 AMD
> and 4 Ivy Bridge). The farm will be increased by 3 Ivy Bridge nodes. We will still
> have the loaner gluon55-gluon57 nodes that we would like to keep around for a little
> while longer until we are routinely running on the AMD machines without issues.
> 
> Let me know if you need any help with this. I realize it is some effort.
> 
> Regards,
> -David
> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/halld-online/attachments/20140929/167087ca/attachment.html