Feature #8157

CSL input to the Fermigrid Bluearc Unmount plan

Added by Arthur Kreymer about 6 years ago. Updated about 4 years ago.

Target version:
Start date:
Due date:
% Done:


Estimated time:
5.00 h
Spent time:
Duration: 24


Liaisons are invited to provide feedback to the
Fermigrid Bluearc Unmount plan found at

Please contact any Task Force member, or contribute to this Issue directly.

We would like to start scheduling ummounts as soon as
individual projects are ready .


#1 Updated by Arthur Kreymer about 6 years ago

Date: Wed, 25 Mar 2015 09:25:57 -0500
From: Gabriel Nathan Perdue <>
To: Arthur E Kreymer <>
Cc: Daniel Ruterbories <>,
Heidi M. Schellman <>

The document and plan look good. Minerva is not ready to unmount bluearc
/data but we have begun testing our reliance on the mount using the new
gpgtest machine you pointed us to. I anticipate being able to unmount on the
timescale of a few months from now if we work at a normal pace on this

#2 Updated by Arthur Kreymer about 6 years ago

Date: Mon, 30 Mar 2015 11:34:04 -0500
From: Michael H Kirby <>
To: Arthur E Kreymer <>
Subject: Fwd: Fermigrid Bluearc Umount plan - request for comments
Here’s some feedback from Herb Greenlee on MicroBooNE.


Begin forwarded message:

From: Herbert Greenlee <>
To: Michael H Kirby <>
Cc: "Eric D. Church" <>
Subject: Re: Fwd: Fermigrid Bluearc Umount plan - request for comments

Hello Mike,

I recommend that the authors insert page numbers.

I don't understand the meaning of "as scheduled" on the Stawman
(page 4).

The report ends section "Background" (bottom page 2) with the
correct statement.

 The only allowed direct access to data has been to Auxiliary highly
 shared files,
for which alternate access methods are now available.

Table on page 9 correctly list microboone as using aux flux files.
seems inconsistent with the small/medium/large categories pages 4/5,
microboone categorized as "small", and small defined as "not using aux
files."  Lar1ND and argoneut are similarly categorized as S+F.
the authors of this report did not consider flux files as fully
as aux files.  Why is that?

The report makes a few mentions of "aux tools being tested" as the
solution to the flux file and aux file problem, but does not give any
details of what these tools are.  I suppose this is meant to include
latest version nutools/GENIEHelper, which supposedly fixes the flux
access issue in a grid-friendly way.

Note that just because a new grid-friendly version of nutools is
as a ups product does not mean that microboone will can start using it
immediately.  Nutools is a compile-time dependency of larsoft.  
Therefore, at a minimum, we need a new version of larsoft.  We need to
know if the migration to the new grid-friendly nutools is fully
transparent, or if it requires, say, changes in fcl files or workflow.
am not personally up to speed on these issues.  I have not seen this
presented in a meeting.  Is there a resource that explains the status,
usage, and impacts of this new nutools/GENIEHelper?

Assuming the new nutools has reasonable compatibility with the current
version, it should be able to propagate fairly quickly into the weekly
larsoft integration release.  We would need the opportunity to test
version in FermiGrid and non-FermiGrid sites.  If we want to continue
using any older versions of larsoft (such as ones being used for
production of mcc6 or other uses), they would have to be patched.
 This is
a fairly large impact, which would be mitigated by having a transition


#3 Updated by Arthur Kreymer about 6 years ago

Version 14, Revision 0.921 03/30

Per feedback from uboone/greenlee

Clarified, changed Strawman 'as scheduled' to 'to schedule'

Moved argoneut, lar1nd, uboone to medium Impact
as they use beam flux files.

#4 Updated by Arthur Kreymer about 6 years ago

I spoke to Rob Kutschke this morning.

Mu2e has some work planned within the next month
to prepare for a June umount.

I will start making Issues to track each project's schedule.

#5 Updated by Tanya Levshina about 4 years ago

  • Status changed from Assigned to Closed

Also available in: Atom PDF