[Condor-users] Recommending OS for new cluster

["Alas, Alex [FEDI]" <aalas@xxxxxxxxxxxxx>]

Guys,

Below is an snapshot of my current condor cluster configuration after running condor_status for reference purposes. We are starting to process some big jobs and the current condor dedicated cluster is not able to process those jobs correctly; the end results are not what they expect, some guy (including myself) in my organization think is related to memory limitations.  When you run those same jobs on local systems that have up to 4 GB of RAM they ran fine but on condor they ran but some of the results are not correct.  Some folks here want to build a new cluster but I am not sure what type of machines are the best suitable to get condor to work. These are the two optional configurations we might acquire but  before suggesting anything I want to consult with everyone here if anybody have any negative experience or anything negative to say  about the RAM or the OS:

 

Option A:  8 Dell Precision R5400 Rack Workstations
(32 cores , 32GB memory)

 

1. PROCESSOR Quad Core Intel® Xeon® Processor E5450 (3.0GHz,2X6M L2,1333)
2. OPERATING SYSTEM Genuine Windows Vista® Business Bonus-Windows XP Professional downgrade
3. MEMORY 4GB, DDR2 SDRAM FBD Memory, 667MHz, ECC (2 DIMMS)
4. OPTICAL DRIVE 8X DVD-ROM with Cyberlink Power DVD™
5. HARD DRIVE 250GB SATA 3.0Gb/s,7200 RPM Hard Drive with 8MB DataBurst Cache™

 

Option B: 4 PowerEdge M600 Blade Servers
(32 cores , 64GB memory)
 
1. PowerEdge M600: Quad Core Intel® Xeon® E5450, 2x6MB Cache, 3.0GHz, 1333MHz FSB ,  2nd Processor:  Quad Core Intel® Xeon® E5450, 2x6MB Cache, 3.0GHz, 1333MHz FSB
2. Memory 16GB 667MHz (4x4GB), Dual Ranked DIMMs
3. Operating System Windows Server® 2003 R2, Standard x64 Edition with SP2,Includes 5 CALs
4. Operating System Media Kits Windows Server® 2003 R2, Standard x64 Edition with SP2, Media Kit

 

Condor_status’ results:

Name	OpSys	Arch	State	Activity	LoadAv	Mem	ActvtyTime
slot1@xxxxxxxxxxx	WINNT51	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot2@xxxxxxxxxxx	WINNT51	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot1@xxxxxxxxxxxx	WINNT51	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot2@xxxxxxxxxxxx	WINNT51	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot1@xxxxxxxxxxxxxxx	WINNT51	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot2@xxxxxxxxxxxxxxx	WINNT51	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot1@xxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot2@xxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot3@xxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot@xxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot1@xxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot2@xxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot3@xxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot4@xxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot1@xxxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot2@xxxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot3@xxxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot4@xxxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
slot1@xxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot2@xxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot3@xxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05
Slot4@xxxxxxxxxxxxxx	WINNT52	INTEL	Unclaimed	Idle	0.000	511	0+02:00:05

           

Total	Total	Owner	Claimed	Unclaimed	Matched	Preempting	Backfill
INTEL/WINNT51	6	0	0	6	0	0	0
INTEL/WINNT52	16	0	0	16	0	0	0
Total	22	0	0	22	0	0	0

    

 

Respectfully,

Alex Alas

Systems Administrator
Fugro EarthData Inc.

Tel. 301-948-8550 x219 Fax 301-963-2064 E-mail: aalas@xxxxxxxxxxxxx

7320 Executive Way, Frederick, MD  21704

Website: http://www.fugroearthdata.com