SupportingHCC

HCC VO

• The Holland Computing Center VO
• Contact: Derek Weitzel <dweitzel@cse.unl.edu>, Adam Caprez <acaprez@cse.unl.edu>; http://hcc.unl.edu/contact/
• Submission system: GlideinWMS

Requirements

• Worker node OS should be RHEL 5, CentOS5 or SL5
• Worker nodes require outbound internet access (nodes can be behind NAT)
• Worker node memory: 1GB / slot minimum, assume 1.5-2 GB
• Worker node scratch: assume 10G per job slot - though this has never exceeded 500 MB
• The worker nodes require the OSG CA certs that are installed as part of the OSG Worker Node Client. Host certs on the worker nodes are not required.
• OSG wn-client is required; lcg-cp heavily used for SRM
• Site squid (optional): used for data 10 MB to 150 MB
• Job preemption is okay
• glexec supported

Security profile

Description:

1. For every job slot, a pilot job process starts up.
2. The pilot job sends outbound TCP traffic over HTTP to a host at UCSD ("factory") and via HTTP to a host at UNL ("frontend").
3. The pilot job spawns a condor startd, which spawns a condor starter.
4. The startd sends outbound UDP traffic to a single port on the frontend. This is randomly chosen (per pilot) from a range of 200 ports currently. This can be changed to TCP if necessary.
5. The starter sends outbound TCP traffic to a port on the HCC frontend. This is chosen randomly (per pilot) from the frontend's ephemeral ports.

Hosts and ports:

• The frontend is glidein.unl.edu (129.93.239.145)
  • ports: 80, 9618-9820
• The factory is glidein-1.t2.ucsd.edu
  • port 8319

Setup

• HCC in MyOSG
• Adding a VO to your site from the OSG documentation, https://twiki.grid.iu.edu/bin/view/ReleaseDocumentation/PreparingComputeElement#Required_User_Accounts
• The steps are generally:
  • Create local UNIX account for the VO (eg. "hcc")
  • In GUMS, create a VO account mapping between the VO and the local UNIX account
  • There is no need to reconfigure the CE - the next time gums-host-cron runs the VO gets added to the osg-user-vo-map.txt file, as well as the osg-supported-vo-list.txt file.
  • You'll need this: vomss://hcc-voms.unl.edu:8443/voms/hcc
  • See GUMS screen shot below

Status

HCC VO status

• HCC VOMS status

Accounting links:

• Generation: http://glidein.unl.edu/glidestats/
• Site summary: http://rcf-gratia.unl.edu/gratia/xml/glidein_hours_bar_smry
• On US ATLAS resources:
  • Since May 2010
  • Since Jan 1, 2011
  • Since May 1, 2011
  • Sliding 14-day view: here

OSG Resources

• Filtered by US ATLAS support center, here

Notation:		completed		work is in progress		table to be updated

Site	OSG Resource	Status	Job Max
TIER1			500
AGLT2			450
MWT2_IU	iut2-grid6.iu.edu:2119		125
MWT2_UC	uct2-grid6.uchicago.edu:2119		500
NET2_BU
UTA_SWT2			10
SWT2_OU_OSCER			64
WT2

Known Issues

US ATLAS - HCC history

• Started on BNL, AGLT2, MWT2_IU, OU_OSCER in 2010
• 20 April 2011: Presentation to US ATLAS Facilities, MinutesApr20#OSG_Opportunistic_Access
• 16 May 2011: report from DW on usage:

  We've now run for ~10 hours.  Right now we're sitting at (for ATLAS sites):
  BNL - 24 slots
  MWT2_IU - 128 slots
  AGLT2 - 233 slots, (note that typical is 50 during full-out ATLAS running, max as stated above)
  The MWT2_IU slots are higher than normal, which seems to indicate they are running preemptively. 

• 17 May 2011: MWT2_UC enabled, 500 slot max, preemptable

GUMS config page example

• Example setting up from GUMS web interface (for sites UC_ITB_PBS and UC_UTB_CONDOR, ibtv-ce-*.uchicago.edu)
• From left menu hit "Shortcut"
• Modifications:
  

-- RobertGardner - 17 May 2011

• HCC usage ending May 18 2011