Skip to content

0.0.17
Compare
Choose a tag to compare
@tmaeno tmaeno released this 31 Jul 18:27
· 4498 commits to master since this release
0.0.17
625a927
  • giving a higher prio to install jobs
  • split runRebro from copyArchived
  • fixed retryInActive to reset file status
  • modified dispatcher to send prodSourceLabel for getJob
  • changed ATLAS_PANDALOG.USERS_ID_SEQ to ATLAS_PANDAMETA.USERS_ID_SEQ
  • added TaskMonitor link to email notifications
  • changed getJob() to allow the prod/analy pilot to get installation jobs
  • fixed retryJobsInActive
  • fixed datasetManager to delete sub from foreign T1 instead of home T1
  • improved getDisInUseForAnal
  • added boostUser
  • improved fairshare to support per-cloud shares
  • changed Setupper to register both DATADISK and PRODDISK as locations for sub
  • changed job/task brokerages not to check DBR with DQ2 at CVMFS sites
  • changed the brokerage to skip release checks for releases=ANY
  • fixed for site.priorityoffset
  • fixed T2 cleanup to check if there is active subscription
  • fixed brokerage and copyArchive for RU
  • changed insertNewJob not to insert metadata when it is empty
  • fixed killUser to kill jobs gradually
  • fixed Setupper to make dis for pin at MCP sites in ND cloud
  • fixed Setupper to take cloudconfig.tier1se into account for dis subscriptions
  • set a limit on G/U in the brokerage
  • sending more info in PD2P logging
  • fixed LFC lookup in the brokerage
  • changed PD2P to be triggered by the second job
  • removed multiCloudFactor from the brokerage for NL
  • added a protection to updateJobStatus to prevent holding->transferring
  • fixed getUserParameter to insert new row if the user is missing
  • fixed Setupper to trigger prestaging when sites with multi-endpoints use TAPE
  • put all info to ErrorDiag in the brokerage
  • added modificationTime constraint to URL sent to the user by Notifier
  • introduced ProcessLimiter
  • changed TA to shorten retry interval after refreshing replica info
  • skipping file availability check for log datasets in TA
  • using cloudconfig.tier1SE to count files at T1
  • setting scope only for ATLAS
  • improved the task brokerage to check datasets with fewer replicas first
  • set limit on the number of IDs to be sent to the logger for reassign/killJobs
  • removed LFC lookup from TA
  • changed PD2P to use secondary share
  • fixed to use correct DQ2 site ID for pinning at sites with multi-endpoints
  • modified to send scopes for output files to the pilot
  • added changeJobPriorities
  • using DATADISK for MCP T1 input at all T1s except US
  • added filespec.scope
  • reducing lifetime of dis when corresponding jobs finished and some of them failed
  • improved the brokerage to count the number of running jobs per processingType
  • using transferringlimit in the brokerage
  • fixed the bulk OK file lookup again for unique ddm endpoint sites
  • reduced interval of PandaMover reattempts to 15min from 3h
  • fixed the bulk OK file lookup in the brokerge for multiple ddm endpoints
  • increased the number of PandaMover channels to 15
  • using DATADISK for MCP T1 input at CERN
  • using a default fareshare defined per cloud if T2 doesn't define share
  • added a protection against overwriting of dataset status by datasetMgr
  • implemented a nested fareshare management mechanism
  • fixed the brokerage message when release is missing for repro
  • fixed TA since replicas at T1 non DATADISK prevented T2 replicas from being used
  • using DATADISK for MCP T1 input at ND,ES,DE,NL,TW
  • added a patch for MWT2 to associate MWT2_DATADISK in TA
  • allowed wildcards in cloudconfig.tier1SE
  • fixed Merger for standalone ROOT
  • fixed Closer to trigger merging for cancelled jobs
  • fixed Setupper to pin DBR as well
  • added a protection to Setupper for file lost after job submission
  • fixed getHighestPrioJobStatPerPG for group queue
  • added group queue to all clouds
  • added FOR UPDATE when getting jobdefID for users
  • removed hard-coded FZK-LCG2_DATATAPE removal in TA
  • set activity=Production to TA subscriptions
  • fixed weight reduction in TA for no input tasks
  • fixed the brokerage to send message to logger for too many transferring's
  • fixed wrong error message in TA when open dataset is incomplete
  • updated TA to use a special weight reduction when only TAPE is available
  • removed selector from fileCallbackListener
  • fixed for TMPDISK
  • fixed Setupper to scan T2 LFC per LFC host instead of per SE
  • fixed Setupper to use correct location when pinning dis at foreign T1
  • fixed sitemapper to allow multiple DQ2 site IDs to use the same token
  • added DQ2 registration time to SLS
  • fixed vomsrenew.sh to check certificate and proxy lifetime
  • fixed file-check in the brokerage for BNL@non-US
  • fixed brokerage not to overwrite file's destSE for destSE=local
  • introduced mcore queue in PG
  • added iscvmfs to SiteSpec
Assets 2
Loading