blog

Franklin was Retired on April 30, 2012

Wed, 09 May 2012 10:55:41 -0700

Franklin was retired on April 30, 2012.

Final Reminder: Please save your $SCRATCH and $SCRATCH2 imporant files by 4/30/12

Fri, 27 Apr 2012 09:29:54 -0700

Franklin batch system is drained, and all batch queues are stopped as of 4/26 23:59pm.

This is the final reminder that please make sure to save important files on your Franklin $SCRATCH and $SCRATCH2. ALL FILES THERE WILL BE DELETED, and there will be no mechanisms to recover any of the files after May 1.

Mon Apr 30: Last day to retrieve files from Franklin scratch file systems
Mon Apr 30, 23:59: User logins are disabled

If you need help or have any concerns, please contact "consult at nersc dot gov".

Franklin retirement date is set: 04/30/2012

Tue, 06 Mar 2012 11:42:58 -0800

The Franklin (and its external login node Freedom) retirement date has been set to April 30, 2012. Below are the related schedules:

Effective immediately: Software frozen except for critical updates
Mon Apr 2: No new accounts will be created
Thurs Apr 26, 23:59: Batch system is drained, batch queues are stopped (no jobs will be running at this point)
Mon Apr 30: Last day to retrieve files from Franklin scratch file systems
Mon Apr 30, 23:59: User logins are disabled

Please note that NERSC will NOT archive any Franklin scratch files. All files on $SCRATCH and $SCRATCH2 will be deleted when the system is retired. Please do not wait until April 30, which is the last day to copy your scratch files to another system (e.g. HPSS). Note that your files in $HOME and /project will be accessible from other NERSC machines after Franklin is retired.

If you are currently only using Franklin, please start migrating to Hopper now. If you need help or have any concerns, please contact NERSC consultants at "consult at nersc dot gov".

Update on Franklin retirement plans

Tue, 21 Feb 2012 09:59:54 -0800

If you are currently only using Franklin you should start migrating to Hopper. If you need help or have any concerns, please contact NERSC consultants at "consult at nersc dot gov".

Franklin to retire no sooner than Apr 30

Tue, 20 Dec 2011 13:19:30 -0800

NERSC is making progress on plans to acquire our next major system. Franklin's retirement is necessary to prepare the machine room for the new system. At this point in our planning we can say that Franklin will retire no sooner than April 30. Additional announcements will be made with more details when a firm date is set.

If you are currently only using Franklin you should start migrating to Hopper. If you need help or have any concerns, please contact NERSC consultants at "consult at nersc dot gov".

Franklin earliest retirement date: 03/31/2012

Tue, 20 Dec 2011 11:54:57 -0800

This is the first announcement for the earliest possible retirement date of Franklin: March 31, 2012. Additional announcements will be made with more details when the date is set.

NERSC is working with DOE on an accelerated schedule for the next major system, and the retirement of Franklin helps us to prepare the machine room for it. If our current plan changes, we may extend the lifetime of Franklin.

If you are currently only using Franklin you should start migrating to Hopper. If you need help or have any concerns, please contact NERSC consultants at "consult at nersc dot gov".

Franklin scheduled maintenance and new software available

Wed, 06 Jul 2011 15:14:35 -0700

In conjunction with the NERSC systems Outages next week (for NERSC Global File Systems maintenance) announced earlier today (see details at: http://www.nersc.gov/REST/announcements/message_text.php?id=2070), Franklin will have a scheduled maintenance on next Wed, July 14, from 8am to 1pm Pacific time.

/project directories will not be available on Franklin from 8am on Wed July 13 to 1pm on Thur July 14 (over 24 hrs). User jobs that need to use the /project file system should add the "gres" directive in their batch job scripts to indicate this requirement. (see detailed instructions at: http://www.nersc.gov/REST/announcements/message_text.php?id=2071)

The following new software versions are available now (as non-default version) on Franklin, and will be set to default at the maintenance.
-- xt-mpich2/5.3.0 and xt-shmem/5.3.0
-- cce/7.4.0

Scheduled maintenance next Wed (June 15) and new software available

Wed, 08 Jun 2011 14:19:46 -0700

1) There will be a scheduled hardware and software maintenance for Franklin next Wednesday, June 15, from 8am to 8pm Pacific time. Please plan your work accordingly and check the NERSC Message of the Day (MOTD) for status update: http://www.nersc.gov/live-status/motd/

2) The following new software versions are available now (as non-default version) on Franklin:
-- pgi/11.3.0
-- perftools/5.2.0 (which includes xt-craypat and apprentice2)

FYI, Issuing the command "module help module_name/version" (e.g., "module help pgi/11.3.0") will display some useful software release information.

If there are no problems reported, the above two new software versions will be set to default on 6/15 at the scheduled system maintenance.

3) Following the maintenance, the per-process CPU limit on Freedom (the external login node of Franklin) will be set to 24 hours (current limit: unlimited). Note: the per-process CPU limit on the internal login nodes of Franklin is 1 hr (remain unchanged).

Batch script keyword for jobs using /project

Thu, 17 Feb 2011 14:48:36 -0800

A batch system keyword has been added to aid jobs that use the /project file system.

#PBS -l gres=project:1

If this keyword is set in a batch script, the job will not be started during a /project scheduled maintance period.

Charge rate changes for reg_big and reg_xbig queues

Mon, 24 Jan 2011 15:08:09 -0800

The charge factor for jobs that run in the reg_big and reg_xbig destination queues has been changed for 0.50 to 0.75 effectitive today.

Wall clock limit increased for reg_small and reg_med queues

Mon, 24 Jan 2011 15:04:34 -0800

The maximum wall clock limit has been increased for the following two destination queues:

reg_small(1-255 nodes): from 24 to 48 hours

reg_med (256-1999 nodes): from 24 to 36 hours.

Jobs can not be submitted directly to these queues. Jobs of the appropriate size and routed automatically from the "regular" submission queue.