Cloud Squad activities

Please see this ADC Reference page for details. This page supplements the instructions in the ADC reference.

References

• ADC Site Exclusion Savannah
• eLog
• ADC shifters email: atlas-adc-adcos-notifications@cernSPAMNOT.ch
• ADC experts email: atlas-adc-expert@cernSPAMNOT.ch

Downtimes

• If scheduled outage, site admins should drain queues 12 - 24 h in advance. Cloud squad should set Panda queues offline (ANALY queue set to test) with eLog and Savannah entries when queues drain.
• If the site outage affects the entire cloud, notify ADC experts to set the cloud offline 12 hours in advance.
• if unscheduled downtime, cloud squad put Panda queue offline (ANALY queues to test) as soon as possible with eLog and Savannah entries.
• When site is ready and before the downtime is declared over, site admins should run siteTests and report to the cloud squad when all jobs are successful.
• When the downtime is declared over and the site admin's tests have passed, cloud squad can then request ADCoS shifters to
  • test and put the Panda production queues online
  • For ANALY queues,
    • Cloud squad (or ADCoS shifters if after hours) set the ANALY queue to "brokeroff" with a comment "HC.Test.Me" so that the Hammercloud tests will run and set the queue online after 5 successful tests.

Responsible for Coverage

Schedule

• from 0900 - 1400 EST, (1300 - 1800 UTC) CA East Coast handle it.
• from 1100 - 1700 PST, (1800 - 0000 UTC) CA West Coast handle it.
• (we will not be covered from 0000-1300 UTC which is significant ...)
• after hours, weekends or holidays: best effort from anyone who sees the issue first.

People

(Please notify your secondary contact if you will be away)

• CA East coast primary: Leslie Groer, secondary: Simon Nderitu
• CA West coast primary: Asoka De Silva, secondary: Di Qing

-- AsokaDeSilva - 2011-05-12