63.2

Summary

Event Service:

- Support for job parameter option --createTAGFileForSE to switch on TAG file creation for event service job (default is off, i.e. the TokenExtractor tool will be used)

- Fix for problem seen in some ES jobs; pilot was not ready to send new event range soon enough after sending previous event range to AthenaMP

Memory monitoring:

- Bug fix for auto-restarts of memory tool by the pilot

- Corrections to location of memory tool on NG

- Added memory tool to exclusion list in looping job killer algorithm

Pilot changes related to leaving the logs where they are produced

- Minor tweaks when GROUPDISK was selected

- For more info see https://its.cern.ch/jira/browse/PRODSYS-504

Version info

Updated the default event index url from ific to cern, in getEventIndexURL() (EventService)

Sending jobReport.json with the metaData field also for event service jobs, in updatePandaServer() (PandaServerClient)

Created useDirectAccess() (Mover)

Sending hasInput to updateRunCommandList() from __main__() (RunJob, RunJobEvent)

Added hasInput argument to updateRunCommandList() (RunJobUtilities)

Removing --createTAGFileForES instruction from runCommandList in updateRunCommandList() (RunJobUtilities)

Updated setUseEventIndex() to set __useEventIndex variable depending on jobPars (RunJobEvent)

Sending jobPars to setUseEventIndex() from __main__(), and moved the call to setUseEventIndex() until job object is defined (RunJobEvent)

Added a waiting loop in listener() for send event range to finish (RunJobEvent)

Added new variables __sending_event_range, __current_event_range and corresponding getters and setters (RunJobEvent)

Now setting __sending_event_range using setSendingEventRange(), and __current_event_range using setCurrentEventRange() in event range loop in \

__main__() (RunJobEvent)

Added memory_monitor to exclusion list for looping job killer in __loopingJobKiller() (Monitor)

Added missing scope_dict and sourceSite in call to sitemover_get_data() in mover_get_data() (Mover)

Corrected self -> runJob in __main__() in event range check (RunJob)

Corrected self -> runJob in __main__() in call to getUtilitySubprocess() (RunJobEvent)

Protection added against no Popen process in executePayload() (RunJob)

Time-out changes:

- verifyCmtsiteCmd() was updated for new timer command output (ATLASExperiment)

- Updated timer command output (TimerCommand)

- Protection added against no Popen process in run() (TimerCommand)

- Converting exitcode to string in case it is not an integer in getProperSiterootAndCmtconfig() (ATLASExperiment)

NG changes:

- Removed obsolete import of timed_command in pUtil

- Updated timedCommand() to use TimerCommand instead of timed_command (pUtil)

- Moved tryint(), splittedname() and isAGreaterOrEqualToB() from PandaServerClient to pUtil

- Now using isAGreaterOrEqualToB() in getUtilityCommand() for the MemoryMonitor setup (ATLASExperiment)

GROUPDISK updates:

- stageOutFile() updated (GFAL2SiteMover, LocalSiteMover)

- put_data() updated (lcgcpSiteMover, lcgcp2SiteMover, curlSiteMover)