Choose timezone
Your profile timezone:
FTS3 - Updates, status and roadmap
- Reminder on SOAP deprecation plans
- LHCb and CMS still relying on it
- No complains in any case with the proposed steps and timing (see slides)
- Resolved to establish periodic FTS3 Steering Meetings
- One slot per month, only if needed
- Open agenda, to be filled as necessary
- Thematic, so parties can know in advance if they are interested
- Need to decide the slot, so it doesn't clash
- Next main topic, probably deployment and configuration
- Gone through Atlas standing issues (PDF attached to the agenda)
- 'Phantom jobs' (job-ids for jobs that failed to commit to the database) are better than 'invisible jobs' (jobs that went to the database, but the client failed to get the job-id)
- FTS3 Dev Team to figure out if it is possible to catch a connection closed from the client
- If so, rollback job if the job-id couldn't be returned
- In any case, there is a prototype implementation in 3.4.0 that allows retrieving non-terminal jobs for a given destination surl
- Duplicated messages for transfers
- Two known causes still there:
- Duplicated transfers (candidate fix in the Pilot@CERN, FTS-336)
- False stalled transfers (still to decide the approach to fix it)
- Two known causes still there:
- Monitoring view grouped by src-dst-activity
- FTS set priority
- Scheduled for 3.4 (FTS-300)
- Recoverable error field to be exposed
- 'Phantom jobs' (job-ids for jobs that failed to commit to the database) are better than 'invisible jobs' (jobs that went to the database, but the client failed to get the job-id)
- Two items brought to attention by LHCb
- "Error 500" when connecting to the REST interface, although the situation is now better (GGUS #117206)
- Reminded that FTS3 service team should announce downtimes in the GOCDB
There are minutes attached to this event.
Show them.
