2.1.1

Alexey Anisenkov:

• Timeout decorator update; preserve/expose original back trace stack on exception in decorated function
• Order replicas by priority (while waiting for Rucio fix); manually sort replicas by priority value since Rucio.list_replicas has a bug in ordering
• Avoid stage out to nucleus; stage out logic updates: ignore the nucleus destination (use a default destination from the astorages). Ported prepare_destinations logic from Pilot 1
• Added new error code 1133, NOSTORAGE

Paul Nilsson:

• First version of Memory Monitor API
• Fix for rucio upload/download timeout issue previously leading to lost heartbeat (problem still exists with rucio command which does not always timeout according to --transfer-timeout option)
• Update to latest detect_cient_location() from rucio with support for IPv6
• Fix for timeout decorator usage in xrdcp copy tool
• Introduced Data API mode data members (requested by A. Anisenkov) instead of passing around mode variable in kwargs
• Added new error code 1334 for AtlasSetup fatal error
• Added support for ALRB_USER_PLATFORM passed from prun/pathena
• Added priority to error code to manually tell the pilot which the most important error is (needed when the job report is processed)
• Updated logic for setting up containers, now supporting ALRB user platform
• Now able to identify out of disk space errors within payload stderr
• Corrected looping job killer (dev Pilot 1 fix adapted to Pilot 2), now finding the latest updated file -before the wrong timestamp was used which led to longer than necessary looping job limits
• Now accepting choices=['SCORE', 'MCORE', 'SCORE_HIMEM', 'MCORE_HIMEM'] in --resource-type pilot option (requested by Fa-Hui Lin)

2.1.0

Large merge with development next branch for the first proper master version that is the first version that runs with real production jobs.

Full version numberL 2.1.0.1