Summary
* Corrected bug in RunJob signal handler. Previously this led to wrong error label (exception). Kill signals handled by the main pilot unaffected
* Updates to aria2c site mover from Vincenzo Lavorini. Bug fix for doubled https://. Using curl option -1 (forces curl to use TLS version 1 when negotiating with a remote TLS server)
* Fix for traces report to Rucio by Wen Guan. Server URL updated
* Time-out fix for S3 object store site mover by Wen Guan
* glExec update from Edward Karavakis to resolve issues at two Canadian sites that do not have the $GLEXEC_LOCATION env var set
* Skipping core count checks for install jobs (job def vs AGIS) to facilitate install jobs on MCORE sites without complicated changes in the installation system
* Updated nightlies setup, fixing a problem with potentially repeated cacheVer strings in the asetup options list (seen with an SLC5 user job)
Postponed until the next pilot version due to illness:
* Sending jobReport JSON file instead of XML for job report versions >= 1.0.0, pending AMI update. Note: only release >=19.2.0 and AtlasDerivation are currently producing the job report v 1.0.0.
Version info
Corrected runJob.getGlobalErrorCode -> runJob.getGlobalErrorCode() in sig2exc(), leading to wrong error code labelling (RunJob, RunJobEvent)
New code from Vincenzo Lavorini; e.g. removed explicit https:// from command line in surls2metalink() (aria2cSiteMover)
Replaced all four occurances of 'https://voatlasrucio-redirect-prod-01.cern.ch' with 'https://rucio-lb-prod.cern.ch' (Mover)
Updated nightlies setup, fixing a problem with potentially repeated cacheVer strings in the asetup options list (seen with an SLC5 user job), in getProperASetup() (ATLASExperiment)
New version of aria2c site mover using -1 curl option. Code from Vincenzo Lavorini (aria2cSiteMover)
Changed time-out from 3600 s to 600 s (xrootdObjectstoreSiteMover)
glExec update from Edward Karavakis to resolve issues at two Canadian sites that do not have the $GLEXEC_LOCATION env var set (pilot, glexec_utils)
Skipping nCore checks for install jobs, in postGetJobActions() (ATLASExperiment)
Getting JSON instead of XML in getPayloadMetadataFilename() (PandaServerClient)
Updated TimerCommand to handle function time outs (Wen Guan)
Added time-outs to S3ObjectstoreSiteMover (Wen Guan)
No comments:
Post a Comment