grid-announce

grid-announce@nikhef.nl

48 discussions

NIKHEF Desktop LCG client tools service interruption
by David Groep 22 Jun '06

22 Jun '06

Dear all, As the filesystem hosting the LCG client tools for desktops at NIKHEF (/global/ices/lcg) is being migrated to the new file server at 10.00 AM, you are kindly requested to refrain from using the client tools from this location between 10.00 and 11.00 AM. Would you also be so kind as to verify that the /global/ices filesystem is no longer mounted on your desktop? Note that job submission from the dedicated User Interface in the NDPF (ui03.nikhef.nl) is *not* affected by this migration and will remain functional throughout the stated period. Regards, DavidG. -- David Groep ** National Institute for Nuclear and High Energy Physics, PDP/Grid group ** ** Room: H1.56 Phone: +31 20 5922179, PObox 41882, NL-1009DB Amsterdam NL **

1 0

scheduled downtime for tbn15.nikhef.nl (se03)
by Ronald Starink 02 May '06

02 May '06

Dear users, On Thursday May 4th, between 9:00 and 17:00, the classic storage element tbn15 (also known as se03) will be unavailable due to an upgrade. During the upgrade, the data on the storage element will not be accessible. If you need to access data currently only available on tbn15, you may want to copy it to the (DPM) storage element tbn18.nikhef.nl before Thursday. Sorry for the inconvenience, Ronald

1 0

tbn20 down again
by David Groep 01 May '06

01 May '06

Dear users, It seems that the current Computing Element hardware (tbn20) has become inherently unstable, so again it is unavailable. You should still be able to submit jobs via our resource broker (rb03.nikhef.nl) from both the centralised user interface as well as your local desktop machines. Sorry for the inconvenience. DavidG. -------- Original Message -------- Subject: Re: [grid-announce] tbn20 down Date: Sat, 29 Apr 2006 22:23:38 +0200 From: David Groep <davidg(a)nikhef.nl> Organization: NIKHEF, Amsterdam To: grid-announce(a)nikhef.nl References: <44532AA9.2060503(a)nikhef.nl> Dear users, Jeff Templon wrote: > our computing element head node, tbn20, is down since last night around > midnight. You should still be able to submit jobs to our resource > broker and they will run elsewhere on the grid. Please report any > problems to grid.sysadmin(a)nikhef.nl. The problems with the NDPF Computing Element have been resolved by manual intervention. As of Sat Apr 29, 22:20 PM CEST the Computing Service should be back to normal operation. Problems can of course still be reported via <grid.sysadmin(a)nikhef.nl> Sorry for the inconvenience caused. DavidG. -- David Groep ** National Institute for Nuclear and High Energy Physics, PDP/Grid group ** ** Room: H1.56 Phone: +31 20 5922179, PObox 41882, NL-1009DB Amsterdam NL ** _______________________________________________ grid-announce mailing list grid-announce(a)nikhef.nl https://mailman.nikhef.nl/cgi-bin/listinfo/grid-announce -- David Groep ** National Institute for Nuclear and High Energy Physics, PDP/Grid group ** ** Room: H1.56 Phone: +31 20 5922179, PObox 41882, NL-1009DB Amsterdam NL **

1 1

tbn20 down
by Jeff Templon 29 Apr '06

29 Apr '06

Hi *, our computing element head node, tbn20, is down since last night around midnight. You should still be able to submit jobs to our resource broker and they will run elsewhere on the grid. Please report any problems to grid.sysadmin(a)nikhef.nl. Have a nice queen's day ... JT

2 1

NIKHEF grid site upgraded; change of service machines
by Ronald Starink 18 Apr '06

18 Apr '06

Dear grid users, The NIKHEF grid site has recently been upgraded to version 2.7.0 of the LCG middleware. During this upgrade, we have installed new machines for the user interface and resource broker services. The new user interface machine is called 'bosui'. It replaces the old user interface host 'tbn12'. The alias 'ui03' now points to the new machine bosui. The user interface software available from the NIKHEF desktops also uses LCG version 2.7.0. The default resource broker is now 'bosheks'. It replaces the machine 'boswachter'. Jobs submitted via the new user interface already use bosheks as resource broker. The output of old jobs submitted via boswachter can still be retrieved. Please retrieve old job output as soon as possible since we will soon stop the services on boswachter. The hosts tbn12 and boswachter will remain in service until April 30, 2006. For news on the NDPF infrastructure, see: http://www.nikhef.nl/pub/projects/grid/gridwiki/index.php/NDPF_News Regards, Ronald -- Ronald Starink ** National Institute for Nuclear and High Energy Physics ** Room: H1.57 Phone: +31 20 5925180 ** PObox 41882, NL-1009DB Amsterdam NL

1 0

[Fwd: WARNING: removal of the LCG RLS Grid catalogue]
by Jeff Templon 19 Dec '05

19 Dec '05

Hi, FYI

1 0

change of NIKHEF resource broker
by Ronald Starink 09 Dec '05

09 Dec '05

Dear grid users, We have changed the resource broker on the NIKHEF-ELPROD facility to the host boswachter.nikhef.nl. Reason for the change is that the previous resource broker bosheks recently suffered from instabilities. For you as a user of the NIKHEF grid facility, nothing changes. Your jobs will automatically be handled by the new resource broker. When submitting jobs, you may notice that the job identifier now contains the hostname boswachter instead of bosheks. Best regards, Ronald -- Ronald Starink ** National Institute for Nuclear and High Energy Physics ** Room: H1.57 Phone: +31 20 5925180 ** PObox 41882, NL-1009DB Amsterdam NL

1 0

quota on sandbox directories
by Ronald Starink 15 Nov '05

15 Nov '05

Dear grid users, The resource broker service in the NIKHEF grid facility has recently been hindered by a full disk. The full disk contained the sandbox directories. (The sandbox directory of a job contains the input files before submission and the output after execution of the job). As a consequence, it was not possible to submit jobs via the NIKHEF grid facility. After investigation of the problem, it was obvious that: 1) Users sometimes store very large files (>> 10MB) in their sandbox directories. 2) The contents of sandbox directories are not retrieved after the job ends. What can you do to prevent this problem? 1) If your jobs produce output files larger than circa 1 MB, do not leave the files in the sandbox directory, but copy the them to a storage element. Please refer to http://www.dutchgrid.nl/Org/Nikhef/tutorial.pdf for an introduction of using storage elements. Smaller files can be left in the sandbox. 2) Fetch the output of your jobs after the jobs have finished. What have we done to prevent this problem in the future? 1) We have enabled quota on the disk containing the sandboxes. Every user has a quotum that is sufficiently large to run hundreds of grid jobs and store the output of small files in the sandbox directories. Any output that does not fit in the quotum will be lost. 2) We will enable clean up scripts that automatically remove the contents of sandbox directories after 28 days. It will not be possible to retrieve the output after this period of 28 days. Best regards, Ronald -- Ronald Starink ** National Institute for Nuclear and High Energy Physics ** Room: H1.57 Phone: +31 20 5925180 ** PObox 41882, NL-1009DB Amsterdam NL

1 0

problems with RB @ NIKHEF
by Jeff Templon 08 Nov '05

08 Nov '05

Hi *, there are problems with the RB at NIKHEF right now. These problems started yesterday afternoon but were 'user-invisible' ... the RB would accept jobs but they were not submitted to sites for execution. I managed to push some of these jobs through earlier today, right now for debugging purposes I have disabled job submission completely. More info as I have it. JT

1 1

problems with NIKHEF cluster
by Jeff Templon 25 Oct '05

25 Oct '05

Hi, For some unknown reason, many of our worker nodes are unavailable. During our attempt to debug the situation we have temporarily disabled the job scheduler. So if you are running jobs directly at NIKHEF (via either qsub or by telling the RB it must go to NIKHEF) your jobs will wait for an as yet unspecified period. Sorry. We think it really sucks too. J "three NIKHEF guys in Pisa trying to fix it" T

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

grid-announce