Difference between revisions of "FrenchSitesInputtoTEGOp"

Un article de lcgwiki.
Jump to: navigation, search
(Input for WG1 : Monitoring & Metrics)
(Input for WG4 - WG 5)
Ligne 23: Ligne 23:
 
* What are the three main problems?
 
* What are the three main problems?
  
===  Input for WG4 - WG 5   ===
+
===  Input for WG4 - WG 5 Mware operational requirements, distribution, deployment, distribution ===
====Scope====
 
* '''Middleware operational requirements'''
 
** Example topics: log files, error messages, manageability, robustness, system requirements, documentation
 
* '''Service and middleware configuration'''
 
** Example topics: puppet, quattor, yaim...
 
* '''Service and middleware deployment'''
 
** Example topics: Packaging, upgrade paths, release schedules, rollout policy
 
* '''Middleware distribution'''
 
**Example topics: Distros - EMI, EPEL, UMD, OSG, repositories, OS support, source availability
 
  
 
==== What is in use currently and works well?====
 
==== What is in use currently and works well?====
* Quattor (FC)
+
* Mware integration using Quattor (long-term support essential)
  
 
==== Top problems====
 
==== Top problems====
* Too many information on logfiles or finding the good information on log file is too hard (IN2P3-LAPP)
+
* Mware usability : Too many information on logfiles. Finding the good information is too hard.
 +
Log management in order to improve Incident Detection and Traceability
  
 
==== What takes the most effort?====
 
==== What takes the most effort?====
* Middelware dependency understanding (IN2P3-LAPP)
+
* understanding the Middelware dependencies
  
 
==== What can be improved about operational procedures?====
 
==== What can be improved about operational procedures?====
* A clear and single channel to announce a new release and documentation (IN2P3-LAPP)
+
* A clear and single channel to announce a new release and documentation
 +
 
 +
==== What can be dropped? ====
  
==== What can be dropped?====
 
  
 
==== Strategic directions you would like to recommend====
 
==== Strategic directions you would like to recommend====
* Middelware compatibilities with the larger set of technical constraints (yaim/quattor, NAS/SAN storage infrastructure,...) (IN2P3-LAPP)
+
* Middelware compatibilities with the larger set of technical constraints (yaim/quattor, NAS/SAN storage infrastructure,...)
  
 
==== Other comments====
 
==== Other comments====
* Aware of EGI stage rollout procedure ? (FC)
+
* Confidence in Mware : EGI staged-rollout on best effort basis. Is this sustainable ?
* A clear statement regarding "EGI staged rollout procedures" is required from WLCG. (FC)
+
* Still needed : relocatable version of Mware
 
+
* Dropped too early : 32 bits UI
 
+
* SL6 WN and disk servers may be needed soon (new hardware procurements)
* besoin relocable mware
+
* Mware distribution via EMI and UMD : no enhancement of interfaces ergonomics
* distribution mware :ergonomie des interfaces EMI UMD
 
* OS : distribution UI en 32 bits toujours nécessaire
 
* solution SL6 pour les workers et serveurs de disque nécessaires rapidement
 

Version du 18:42, 10 novembre 2011

BACK to TEG

Input provided by french sites

Input for WG1 : Monitoring & Metrics

Input for WG2 : Support tools & Underlying services + WLCG operations

  • Scope ====
    • Support tools : Ticketing, Accounting tools, Request trackers,Administration tools [GOCDB, CIC]
  • Underlying services : Messaging services, Information services
  • WG2 Sub-Wiki : https://twiki.cern.ch/twiki/bin/view/LCG/WLCGTegOperationsWG2
  • What works well? What is used ?
  • What are the three main problems?

Input for WG3 : Application Software Management

Input for WG4 - WG 5 Mware operational requirements, distribution, deployment, distribution

What is in use currently and works well?

  • Mware integration using Quattor (long-term support essential)

Top problems

  • Mware usability : Too many information on logfiles. Finding the good information is too hard.

Log management in order to improve Incident Detection and Traceability

What takes the most effort?

  • understanding the Middelware dependencies

What can be improved about operational procedures?

  • A clear and single channel to announce a new release and documentation

What can be dropped?

Strategic directions you would like to recommend

  • Middelware compatibilities with the larger set of technical constraints (yaim/quattor, NAS/SAN storage infrastructure,...)

Other comments

  • Confidence in Mware : EGI staged-rollout on best effort basis. Is this sustainable ?
  • Still needed : relocatable version of Mware
  • Dropped too early : 32 bits UI
  • SL6 WN and disk servers may be needed soon (new hardware procurements)
  • Mware distribution via EMI and UMD : no enhancement of interfaces ergonomics