support for system administrators feedback from site admins after testbed1 experience 05/03/2002 4th...
TRANSCRIPT
Support for system administrators
Feedback from site admins after Testbed1 experience
05/03/2002 4th EDG WS ParisG. Merino, IFAE
TB1 integration from site admins
• We (spanish Testbed sites) are one of the EU-DataGrid “remote” (as described in D6.1) Testbed sites– Not yet in the Testbed1
• Here, I will briefly report on our activities on the Testbed1up to now and try to extract some “Testbed site needs” from this experience
Status of the spanish Testbed sites
IFAE (Barcelona) - AC, Proj. coord.
Status of the spanish Testbed sites
IFAE (Barcelona) - AC, Proj. coord.
IFCA (Santander) - Spanish CA
IFIC (Valencia)
CIEMAT (Madrid)
Funded effort distribution for the 1st year
Status of the spanish Testbed sites
IFAE (Barcelona) - AC, Proj. coord.
IFCA (Santander) - Spanish CA
IFIC (Valencia)
CIEMAT (Madrid)
UAM (Madrid)
UNIOVI (Oviedo)
Funded effort distribution for the 1st year
Unfunded effort. Already some machines connected to the TB
Status of the spanish Testbed sites
IFAE (Barcelona) - AC, Proj. coord.
IFCA (Santander) - Spanish CA
IFIC (Valencia)
CIEMAT (Madrid)
UAM (Madrid)
UNIOVI (Oviedo)
UB (Barcelona)
USC (Santiago)
Funded effort distribution for the 1st year
Unfunded effort. Already some machines connected to the TB
Getting involved (CrossGrid, National Grid initiatives…)
EDG TB1 s/w installation status
• Our 1st target: – Install the 2 basic Grid Elements on each site
CE
SE
WNGatekeeper +
GDMP server(s) Alice,Atlas,CMS,LHCb
Biomed
Eathobs
Wpsix, iteam
EDG TB1 configuration @IFAE
grid-s1.ifae.es• GDMP-Atlas VO (for the moment)
• Gatekeeper (fork)• NFS server, exporting:
/home/atlas001,…
/etc/grid-security/certificates
(also grid-mapfile and gridmapdir)
• edg-crl-update• edg-mkgridmap• edg-pinger
WNGatekeeper +
GDMP serversAtlas
grid-w1.ifae.es
• Gatekeeper• PBS• GIIS
– ifae (local site GIIS)– es (country GIIS)
Testbed1 activityTry to “stay tuned” within the information flow:
– Documentation • Key document – EDG Installation Guide• Other documents in the WP6 Web page
– Bugzilla• Keep an eye on it to be aware of bug status• Use it to report about malfunction discovery
– Mailing lists, specially…• www.listbox.cern.ch/earchive/hep-proj-grid-integration-team/
+
Rea
l-tim
e in
fo
-
Testbed1 activityTry to “stay tuned” within the information flow:– Documentation
• Key document – EDG Installation Guide• Other documents in the WP6 Web page
– Bugzilla• Keep an eye on it to be aware of bug status• Use it to report about malfunction discovery
– Mailing lists, specially…• www.listbox.cern.ch/earchive/hep-proj-grid-integration-team/
General comment:Up to now it has been quite hard for a “remote site” to follow the Testbed1 evolution from those sources
“We are shooting on a moving target” (F. Gagliardi)
TB info sources: EDG Install Guide
The EDG Installation Guide
• Main source of information• It is an “evolving document” (current from 3/02/2002)
– It is “condemned” to be incompletehttp://www.pi.infn.it/~flavia/se_config.html
http://www.lnl.infn.it/datagrid/wp4-install/testbed-report_2/index.html
• Sometimes the info is a bit confusing or inconsistent w.r.t. other WP’s docs– E.g. Configuration of the CE & SE Info Systems…
TB info sources: EDG Install GuideE.g. Configuration of the CE&SE Info Systems:
• Examples of info-mds.conf are given twice– 7.5(7.6) CE(SE) Configuration, Configuring GRAM and GRIS– 8.1 Ftree&MDS Info Services and Info Providers
• For the parameters inside globus.conf:– From WP3 Web (http://hepwww.rl.ac.uk/DataGridMonitoring)
GRID_INFO_GRIS_REG_GIIS=ral - The site name GRID_INFO_GRIS_REG_HOST=hostname.rl.ac.uk
– From WP3 Document “MDS Deployment Testbed-1”GRID_INFO_GIIS_1=tb1-pbs GRID_INFO_REG_GIIS_1=ralGRID_INFO_REG_HOST_1=hostname.rl.ac.uk
– From EDG Installation GuideGRID_INFO_GIIS_1=ce - The GIIS name GRID_INFO_REG_GIIS=ral - The site nameGRID_INFO__REG_HOST=hostname.rl.ac.uk
TB info sources: Bugzilla & mailing list
Bugzilla (http://marianne.in2p3.fr/datagrid/bugzilla)
• Keeps record of s/w features/bugs• Information about which are the open/closed
problems at a given moment• Searchable: Clear classification in terms of
programs (GDMP, LCFG…) and versions• New category “Testbed configuration” added on
mid December:– This might be the best info repository for site admin
issues– Still, some information here is not totally up to date…
TB info sources: Bugzilla & mailing list
The ITeam mailing listwww.listbox.cern.ch/earchive/hep-proj-grid-integration-team/
• This is the ultimate source of truly real-time information concerning TB1 installation issues– E.g. Replica Catalog installation instructions,
SE/GDMP configuration issues…
• The throughput is high enough (~102 mails/day) to make online reading/filtering a tough task
• Searchable: But not as good as bugzilla to search for “all those problems related to a given service”
MapCenter
• Online information about LDAP based Information Systems
• User friendly browsable
Countries 2002/03/05 08:58:30 GMT (refresh=5min)
Map Link Symbol No Status Normal TCP failed Ping failed
Czech_Republic Geographical List->Czech Republic
Denmark Geographical List->Denmark
Finland Geographical List->Finland
France Geographical List->France
Germany Geographical List->Germany
Ireland Geographical List->Ireland
Italy Geographical List->Italy
Netherlands Geographical List->Netherlands
Norway Geographical List->Norway
Portugal Geographical List->Portugal
Russia Geographical List->Russia
Spain Geographical List->Spain
Sweden Geographical List->Sweden
Switzerland Geographical List->Switzerland
United Kingdom Geographical List->United Kingdom
More info sources: WP6 secure Web
http://marianne.in2p3.fr/
Extensive info on real
TB machines config.
rpm -qa/ps -ef output
.conf files
…
1. Machine Name
lxshare0219
2. Grid role(s) WN site 1.
3. Installed Software Packages
3.1. Soft Pack n.1
3.2. Soft Pack n.2
4. Relevant Install Path (Package)
4.1. /home/path1/...(Package n.1)
4.2. /etc/path2/... (Package n.2)
5. Used Network Services
5.1. Network Service N. (port)
5.2. network Service N. (port)
6. Used Service Port (Service)
6.1. Port# n.1 (Service n.1)
6.2. Port# n.2 (Service n.2)
7. Configuration files (Service)
7.1. Config FileName 1 (Service n.1)
7.2. Config FileName 2 (Service n.2) 8. rpm -qa - HyperLink to the command output -
9. Running Deamons
9.1. Deamon n.1
9.2. Deamon n.2 10. ps -efl - HyperLink to the command output -
11. Relevant info for the user (JDL file)
11.1. .....
11.2. ..... 12. /etc/services Hyperlink to the file
13. /etc/inetd.conf Hyperlink to the file
14. Comments ...
You are /C=ES/O=DATAGRID-ES/O=IFAE/CN=Gonzalo Merino Switch to HTTP . Website Help. Built with GridSite 0.1.3
More info sources: National WP6 Web sites
• All them accessible from marianne.in2p3.fr
• Several (Dutchgrid, Nordugrid, GridPP,…) include useful information on the EDG s/w installation & configuration:– Step-by-step instructions for installing services– Comments on the “official” installation procedures– …– Some really fancy monitoring information…
More info sources: MDS monitoring
MDS single-host response times from dutchgrid.nl web
Summary (what we need/have from WP6?)
Bugzilla “Testbed Configuration” categorySite administrators mailing listWP7 monitoring tools such as MapCenter
Configuration of some “reference” machines (CERN?) available from a secure Web siteEDG Installation Guide: collect all the useful information that is dispersedUseful info on national WP6 web sites (tools, step-by-step installations…) could be compiled somewhere Planning & schedule for widespread TB deployment to remote sites