Minutes for Regional Operations Meeting DECH (October 6th, 2006)
Attendance:
Clemens Koerdt <chair>, Guenther Grein (FZK)
Christoph Wissing (Uni Dortmund, too), Uwe Ensslin, Andreas Gellrich
(DESY-HH)
Christian Peter (ITWM)
Hans-Guenther Borrmann (Uni Freiburg)
Horst Schwichtenberg (SCAI)
Peter Kunszt (CSCS)
Apologies*:
Kilian Schwarz (GSI)
Torsten Harenberg (Uni Wuppertal)
Missing:
(DESY-ZN)
(RWTH-Aachen)
(MPPMU)
(Uni Karlsruhe)
1. Introduction
Announcements:
new old report cycle: Operations meeting now back to Mondays,
accordingly report window will not change as announced in the last meeting.
SL4: posted link to talk from Markus Schulz (see agenda). so far 90% of
gLite ported to SLC4 to be finished by end of October. Deployment
expected before the end of the year. Please start planning for the
migration. perhaps start with pps and gather experience with the
different dependencies.
VOMS: one of the services that will be pushed for towards the end of the
year. See also WLCG Commisioning schedule (linked to the agenda page).
There are also links on talks concerning the deployment schedule and
yaim installation (see agenda page).
Job priorities: talk of Jeff T. linked to the agenda page: experiments
will use voms roles for sgm and normal users, make sure to have voms
configured correctly.
Collection of items to be forwarded to the TCG:
use this forum to collect and discuss issues to be raised. Horst will
gather items and prepares a report on a regular basis
Actionitems
- Action point: posted network measurement survey -> so far no answers
received. please participate ! Open.
- communication channels: complicated situation, wants everyone to
decide if he wants to receive mails (admin list or personal mail
addresses) -> communicate your preferences!Open.
- Assessgrid: sites are still encouraged to participate -> they are
interested in job traces, hardware information (like CPU temperature,
fan speed) and node maintenance cycles. Please participate!Open.
- Dech VO: Supported? (Ute has made tests --> to be
submitted, nothing much seems to have changed though). Open.
- VOs: Clemens still to send list around (!#) -> everything is in the
CIC portal, sites see no particular need for a detailed list. Closed.
- SFT-Server: installed, Apache, Tomcat and MySQL running, one servlet
missing that should listen on port 8088 related to RGMA Server
functionality... -> Open.
2. Round the Sites
DESY
----
started to setup PPS. again no easy thing. usual problems with missing
dependencies. newly installed LFC on PPS not working, Workload manager
installation also not straightforward.
Production: heavily participating in CMS data challenges. annoyed by
changed fetch CRLs for CA certificate updates, lost jobs because of
that. Annoyed by hardcoded java version _08 in the m/w.
Also opened bug concerning the draining of the RB. Answer was
unsatisfactory: blocking port 9002 -> user will see native api error and
eventually change rb, which is a bad concept
C.K.: this is part of the issues that will be forwarded to the TCG.
FZK
----
Experiencing various routing issues, that led to a short network outage and failing rm tests.
Fixed RB instability (was out of disk space)
planning to update dCache in two weeks time to 1.6.6.6
SCAI
----
heavily involved in biomed data challenge, disc with file system crashed
after heavy usage from biomed, no jobs running for some time due to
that, problem with biomed licence server
experienced the known openssh upgrade problems, lost biomed docking data
because of that!
biomed will use WMS starting from november
trying to produce a glite standalone version for a training course in
paris. experienced problems with setting up the WMS, strange thing
because of existing working WMS at SCAI itself, related to gLite version
3.0.2? or simply some paths set incorrectly?
CSCS
-----
no bigger problems. experience the troubles with the openssh update that
lead to JS failures.
expects new storage delivery and will then change from DPM to dCache,
could probably ask for support then
cscs heavily involved in cms data challenge with good results
would like to use dech VO temporarily for new application: core grid
(only one user, probably for December, does not need resources)
C.K.: this is also what dech VO was designed for, see no problems with this
Dortmund
--------
everythings working fine apart from a problem with failing LHCb data
transfers. Investigating.
Freiburg
------
experiencing some shortage of personel. Peter Wienemann has left.
Guenther only available for about half a working day. local VO seem to
work fine. some job problems with LHC experiment. investigating
ITWM
----
new hardware has arrived, SE extended and new nodes included into the
production system
informations system experienced heavy load, will be moved next week, SD
is planned
somewhat surprised that new machines did not need SL4, they are working
fine so far under SL3
3. COD
Have given feedback at last COD meeting in Geneva: was well received.
Integrated various working groups there. Christian was working with
Italian team concerning failover.
4. ROC-On-Duty
G.G.: There are still quite some old tickets, one should check wether
they can be closed.
DESY support group now split into two seperate ones: DESY-HH and DESY-ZN.
H.S.: no further questions concerning handover
AOB
---
none
There are minutes attached to this event.
Show them.