CERN Accelerating science

All releases

FTS-Rest 3.4.1

Release Notes - FTS - Version fts-rest 3.4.1

Bug

  • [FTS-430] - Explicitly request application/json on the configuration web UI
  • [FTS-441] - Use "READ COMMITTED" isolation level
  • [FTS-443] - Make sure transactions are closed when the session falls out of scope

New Feature

  • [FTS-372] - Namespaced uuid version 5 for Atlas

FTS-Rest 3.4.0

Release Notes - FTS - Version fts-rest 3.4.0

Bug

  • [FTS-269] - /jobs/<job-id>/files leaks db connections
  • [FTS-270] - IntegrityError on submission when another job inserted the same storage pair in t_optimizer_active
  • [FTS-282] - REST CLI: -K is a flag, but it is expecting a value
  • [FTS-291] - Forbid reuse jobs with different hosts
  • [FTS-296] - Use SHA256 for signing certificate request
  • [FTS-297] - When un-banning a storage, need to set wait_timestamp to Null
  • [FTS-307] - Delegation on Oracle fails with a 500 when there is an existing credential in the cache
  • [FTS-341] - voms_attrs empty for voms proxies delegated to FTS
  • [FTS-342] - When using RFC proxies in http authentication the user_dn is incorrect
  • [FTS-358] - FTS REST APIs do not support alternative ca certificates locations
  • [FTS-394] - Connection validation must be called on checkout

Configuration Change

  • [FTS-289] - Allow bring online jobs for mock://

Improvement

  • [FTS-247] - Use sqlalchemy 0.8 available in epel6
  • [FTS-254] - Allow to enable IPv6 at submission time
  • [FTS-274] - Add --capath to the fts rest CLI
  • [FTS-293] - Expose deletion jobs in the easy API
  • [FTS-299] - REST Client: If X509_USER_PROXY is not present, try with the default location
  • [FTS-300] - Allow to modify a job priority while it is on the queue
  • [FTS-301] - Add a tool that evaluates if a host is still running something
  • [FTS-312] - When accessing with the host cert, grant full permissions
  • [FTS-405] - Migrate from pycurl to 'requests'

Incident

  • [FTS-365] - Status of ATLAS Canceled jobs @RAL and @CERN cannot be queried via REST

New Feature

  • [FTS-221] - Allow to retrieve files/jobs with a given destination surl
  • [FTS-234] - Small tool that queries monitoring to see active limits
  • [FTS-272] - Activity shares missing in the REST config API
  • [FTS-290] - Expose retry_delay on submission
  • [FTS-316] - New algorithms to select best replica at submission time
  • [FTS-321] - REST CLI: Use hostcert.pem/hostkey.pem and localhost if no parameters are specified

Task

  • [FTS-266] - Add FTS dropbox/s3 configuration to FTS Web Management interface
  • [FTS-268] - Enable by default ExportCertData
  • [FTS-317] - Add order by descending finish_time when showing jobs and a limit is specified
  • [FTS-320] - Extend HTTP authorization to accept proxy chain
  • [FTS-356] - need REST equivalent of gSOAP ifce cancelAll

fts 3.3.3

Bug

  • [FTS-326] - In the 'backup' method, nFiles and nJobs are not modified
  • [FTS-336] - The 'update status' check is done after the url copy had already been spawned
  • [FTS-339] - When a session reuse job with retry enabled has *some* transfers going to be retried, they starve
  • [FTS-340] - Strict-copy set by job, overriden by link configuration
  • [FTS-351] - fts_url_copy ignores --debug

Improvement

  • [FTS-338] - Database backup method uses way too big transactions sometimes, allow to configure
  • [FTS-352] - Introduce back check for available memory left before spawning fts_url_copy

FTS 3.4.2

fts 3.4.2

Bug

  • [FTS-232] - Redundant where statement
  • [FTS-235] - Some transfers may have the transfer host set, but no log exists
  • [FTS-241] - Transfers executed twice when the node pool changes
  • [FTS-273] - URL Copy sends two messages for the same file if signals are involved
  • [FTS-292] - Multihop transfers must send a message per hop
  • [FTS-294] - Priority seems to be broken
  • [FTS-311] - max_time_in_queue value is ignored
  • [FTS-314] - Option checking not working as intended in some fts command line clients
  • [FTS-326] - In the 'backup' method, nFiles and nJobs are not modified
  • [FTS-328] - Hosts still marked as 'drained' are neved removed from t_hosts
  • [FTS-329] - Need to remove <availability> from the sls output
  • [FTS-331] - Debug should be disabled by default on the monitoring
  • [FTS-332] - Web monitoring leaks memory
  • [FTS-336] - The 'update status' check is done after the url copy had already been spawned
  • [FTS-339] - When a session reuse job with retry enabled has *some* transfers going to be retried, they starve
  • [FTS-340] - Strict-copy set by job, overriden by link configuration
  • [FTS-351] - fts_url_copy ignores --debug
  • [FTS-371] - Activity sharing tags do not correspond between the client and the monitoring
  • [FTS-379] - Fix the Web monitoring for CentOS7
  • [FTS-380] - Web monitoring: escape urls on the query args
  • [FTS-401] - Segfault when AuthorizedVOs is empty or missing

Improvement

  • [FTS-227] - Add support for wildcard in the operation limit
  • [FTS-237] - Introduce a limit of monitoring messages to be read and sent from the disk
  • [FTS-238] - Log tranfers to purge and logs to rotate to different directories
  • [FTS-254] - Allow to enable IPv6 at submission time
  • [FTS-281] - Expose recoverable/non recoverable via messaging
  • [FTS-298] - Monitoring: Allow to set page size
  • [FTS-303] - Remove 'queued' from server statistics
  • [FTS-305] - Store in t_optimizer_evolution entries also for links with fixed numbers
  • [FTS-309] - Reduce bring online bulk request to 1k
  • [FTS-325] - SLS now expects a 'status' field
  • [FTS-333] - Web monitoring in a separate wsgi process
  • [FTS-334] - Web Mon: Use MySQLdb.cursors.SSCursor to reduce memory footprint
  • [FTS-335] - Use system's coredump capability rather than Google Coredumper
  • [FTS-338] - Database backup method uses way too big transactions sometimes, allow to configure
  • [FTS-352] - Introduce back check for available memory left before spawning fts_url_copy
  • [FTS-375] - Systemd ini scripts required
  • [FTS-377] - Operation limit missing in the monitoring
  • [FTS-389] - Make 'stalled transfer' check configurable
  • [FTS-396] - Make RAM requirements for auto-drain configurable

New Feature

  • [FTS-288] - Expose file size in transfer view
  • [FTS-306] - Show fixed values configurations
  • [FTS-315] - Change some cli clients to use rest by default
  • [FTS-369] - Overview including activities
  • [FTS-402] - Re-enable AuthorizedVO

Task

  • [FTS-283] - Remove ftsdb Python wrappers
  • [FTS-285] - Deprecate SOAP Python bindings
  • [FTS-287] - Remove MonitoringDb code

WebFTS 2.2.6

Bug

  • [FTS-409] - Cannot resubmit transfers from/to dropbox
  • [FTS-410] - Endpoint content is not ordered

WebFTS 2.2.5

Bug

  • [FTS-361] - When switching between tabs sometimes the left SE endpoint is not saved
  • [FTS-378] - Cannot list dav+3rd protocol via REST API

Task

  • [FTS-362] - Add support contact to webfts.cern.ch main page

WebFTS 2.2.3

Bug

  • [FTS-264] - Wrong file Attributes displayed for Grid SE endpoints

FTS 3.3.x

FTS3 3.3.2

Bug

  • [FTS-242] - SLS now requires namespaces
  • [FTS-246] - If hostcert.pem and hostkey.pem are owned by fts3, Globus loads them first
  • [FTS-253] - Checksum errors are *not* recoverable
  • [FTS-256] - Proxy expired should be NON recoverable
  • [FTS-257] - If the connection to the broker go away, should be able to recover
  • [FTS-259] - --proxy is only used for REST
  • [FTS-241] - Transfers executed twice when the node pool changes (Reported by ATLAS, affects all previous releases)
  • [FTS-273] - URL Copy sends two messages for the same file if signals are involved (Reported by ATLAS, affects all previous releases)
  • [FTS-331] - Debug should be disabled by default on the monitoring

Improvement

  • [FTS-224] - Log by default if the destination file has been deleted or not
  • [FTS-225] - Generate coredump in addition to the backtrace
  • [FTS-243] - Missing fields in the server configuration view
  • [FTS-258] - Log if the error is considered recoverable or not recoverable
  • [FTS-260] - Messages: Timestamp should have milliseconds resolution
  • [FTS-333] - Web monitoring in a separate wsgi process
  • [FTS-334] - Web Mon: Use MySQLdb.cursors.SSCursor to reduce memory footprint

New Feature

  • [FTS-244] - Allow to specify the FTS3 endpoint via an environment variable
  • [FTS-245] - Use gfal2 functionality to pass some extra information to the storages

Task

  • [FTS-248] - Prepare packaging to reduce dependency with Bamboo

FTS3 REST 3.3.3

Bug

  • [FTS-208] - Reduce memory consumption
  • [FTS-209] - Avoid updates of t_optimize_active on submission
  • [FTS-212] - In credentials.py, we are binding from a dicitonary without copying, which has side effects
  • [FTS-261] - REST Client: Context seems to be leaking (Reported by LHCb, affect all previous releases)
  • [FTS-265] - When polling multiple jobs with file fields at once, inner fields are duplicated between jobs (Reported by ATLAS, affects only 3.3.0)
  • [FTS-269] - /jobs/<job-id>/files leaks db connections (Found by Andrea Manzi, affects previous versions of 3.3)
  • [FTS-270] - IntegrityError on submission when another job inserted the same storage pair in t_optimizer_active (Reported by ATLAS, affects latest 3.2.x)

Improvement

  • [FTS-205] - Allow specifying the priority on submission time
  • [FTS-207] - Improve performance of job listing
  • [FTS-210] - Improve OAuth2 granularity
  • [FTS-211] - Allow time_window to be in minutes as well
  • [FTS-214] - Select the first best replica at submission time via REST
  • [FTS-215] - Forbid submission of multiple replica combined with other types of jobs
  • [FTS-216] - Re-factorize a bit, paying attention to SonarQube output
  • [FTS-222] - Try to reduce job submission latency
  • [FTS-223] - Validate version as part of the packaging

New Feature

  • [FTS-183] - Web interface to use the REST configuration API
  • [FTS-203] - Requested: Allow to specify only a subset of the job's fields
  • [FTS-204] - Add support for urn:ietf:wg:oauth:2.0:oob
  • [FTS-217] - Allow to cancel a single (or set of) file within a job
  • [FTS-218] - Expose if an error is recoverable or not recoverable via REST API
  • [FTS-219] - max_time_in_queue is missing in REST

Task

  • [FTS-248] - Prepare packaging to reduce dependency with Bamboo

FTS 3.2.33

This is mostly a bug-fix release, with particularly important improvements on the REST API.

FTS3 Release Notes

FTS3-Rest Release Notes

 

FTS 3.2.32

  • Jobs priority can now be set at submission time (via REST only)
  • Activity shares fixes
  • Monitoring using messaging fixes to ensure that no messages are being lost

Detailed release notes:

https://svnweb.cern.ch/trac/fts3/milestone/FTS3%203.2.32

Pages