Heart beat and health

1
Heart beat and health
2
2 axis

• Me versus joe user
• is MY job going to run on MY resources right now?
• Simple versus complex
• On/off versus performance and more complex

3
Tool Information

• Schema
• Central vs dist control
• User vs root
• Permissions ( account)
• What do you test
• Archiving
• Client

4
Inca (TeraGrid)

• Schema Glue based XML (but some alteraations)
• Central control
• User level
• Generic inca account
• What do you test
• Full TG SW stack, GridServices, Cross site tests
• Archiving - yes
• Client several web views

5
GITS (Southampton, Manchester)

• Schema
• XML home grown
• Central vs dist control
• Local site control, cron job
• User vs root
• Well defined user with a cert
• New server cert
• What do you test
• Globus GT2 and WS-RF based service tests
• Pt to pt authentication, job sub, file transfer,
  MDS, gass, grid ftp , gsi ssh,
• Archived - yes
• Client web w/ red and green boxes

6
Jim Basney Scripts (UIUC)

• Schema
• XML home grown
• Central control
• Local tests only
• User account
• What do you test
• GT2, MPI
• Archiving - no
• Client web page

7
Service monitoring

• Many others that are roughly equivalent to these

8
GridMon, Pinger, mapcentre (egee)

• Network monitors (only)
• Several levels of nw tests
• Root
• Schema couple are use ggf nm schema
• Central control
• Archiving
• Client views often web

9
GeoDISE user level

• User Simon (run by user)
• Central vs dist control by user
• What do you test
• Authentication with a specific resource (timed)
• Test job on grid resource (perf chars received)
• File transfer to and from that resource
• Executable existence
• Working directory exists (user defined)
• Schema - internal
• Archiving
• Client

10
Things to know

• 1) taxonomy
• 2) service level inca incorporating gits tests
  possible
• 3) nw still many
• 4) schemas service data has no standard, nw is
  in ggf (nm-wg)
• 5) Difference between info collected and how it
  is used (client views), but many similar tests