v9.1.7

[integration]

*Docs

FIX: (#8519) added NumberOfGPUs and MemoryLimitMB to the configuration example

*Resources

FIX: (#8518) handle NoValidConnectionsError properly in SSHCE

*WorkloadManagement

FIX: (#8512) move setCredentials back to its original location in SD

*Core

FIX: (#8510) Improved error message. Fixes #8509
FIX: (#8507) remove MySQL ping reconnect
FIX: (#8491) Improve VOMS2CSAgent to handle nested groups and only sync groups that have the voms.role label

*FrameworkSystem

FIX: (#8503) getCAs and CRLs unsetting X509_CERT_DIR

v9.1.6

[integration]

v9.1.5

[integration]

*FrameworkSystem

FIX: (#8499) delete proxy does not need a group

*WorkloadManagementSystem

FIX: (#8495) moved getJobParameters in JobParametersDB

v9.1.4

[integration]

*FrameworkSystem

FIX: (#8493) correct order for voms-sync cli

*Core

FIX: (#8490) load the MonitoringReporter only if needed
NEW: (#8484) Add DIRAC_FAST_PROCESS_POOL as experimental feature to speed up the REA

*WorkloadManagement

CHANGE: (#8479) get cpu work loeft from a single source of truth

v9.1.3

[integration]

*CI

CHANGE: (#8488) Replace --diracx-dist-dir with --diracx-src-dir

*TransformationSystem

CHANGE: (#8486) factorize building a Request from a transBody json

*WorkloadManagementSystem

FIX: (#8481) SiteDirector._setCredentials should renew the proxy only if below the queue CPUTime

v9.1.2

[integration]

*ResourceStatusSystem

FIX: (#8480) restore CEs in getSiteElements

*Github actions

FIX: (#8478) Reenable automatc release deployment in CVMFS

*WorkloadManagementSystem

FIX: (#8477) pass the job ID to the dirac-cwl job wrapper

*MonitoringSystem

FIX: (#8363) MonitoringReporter - avoid endless loop in sending records with failures

v9.1.1

Deployment

MySQL changes (optional, but recommended)

USE ResourceManagementDB;
DROP TABLE AccountingCache;

CS changes (optional, but recommended)

remove CS entry /Operations/<>/ResourceStatus/StatusTypes

Release notes

*ResourceStatusSystem

CHANGE: (#8473) removed ResourceManagement.AccountingCache table and methods
CHANGE: (#8473) removed Catalogs from list of managed RSS resources

*Interfaces

CHANGE: (#8470) added includeInSandbox flag to Job().setExecutable(). Default: True
FIX: (#8470) Job().setExecutable(): warn->verbose message when the executable is not local

v9.1.0

Deployment

MySQL changes (optional, but recommended)

USE JobDB;
ALTER TABLE Jobs ADD INDEX JobName (JobName), ALGORITHM=INPLACE, LOCK=NONE;

USE PilotAgentsDB;
ALTER Table PilotAgents ADD INDEX idx_dest_queue_status (DestinationSite,Queue,Status), ALGORITHM=INPLACE, LOCK=NONE;

USE ResourceStatusDB;
DROP TABLE NodeHistory;
DROP TABLE NodeLog;
DROP TABLE NodeStatus;

Release notes

*FileCatalog

FIX: (#8467) RucioFileCatalogClient.setMetadata returns empty result on success #8330

*DataManagementSystem

CHANGE: (#8466) FTS3 jobs don't call delete when TPC fails

*WorkloadManagement

NEW: (#8464) add UseDiracXBackendOnly CS option to SandboxStore to make DiracX the exclusive sandbox backend and disable local purging
FIX: (#8463) Add composite (DestinationSite, Queue, Status) index to PilotAgents table for SiteDirector.countPilots()
FIX: (#8463) Add JobName index to Jobs table for TransformationSystem and monitoring lookups
FIX: (#8462) Optimize TaskQueueDB matching queries by adding composite (TQId, Value) indices and replacing COUNT subqueries with EXISTS/NOT EXISTS

*docs

NEW: (#8446) added RSS technical documentation

*ResourceStatusSystem

CHANGE: (#8446) Removed nodes (queues) tables

v9.0.20

[integration]

*WorkloadManagement

FIX: (#8457) cache SandboxMetadataDB instance in JobSanity and JobCleaningAgent to avoid unnecessary DB overhead per job

v9.0.19

[integration]

*TransformationSystem

FIX: (#8452) Bulk TransformationDB.getFileIDsForLFN to avoid full in memory table

*Tests

NEW: (#8420) script to debug and test CEs and the DIRAC interfaces to interact with them

*RucioFileCatalog

fix: add missing option in RucioFileCatalog.getUserMetadataBulk