A/L morning.
Meetings in morning.
Worked on rucio-analysis package, expanding on Rohini's non-deterministic ingestion test to make the test more flexible to changes in the PFN schema.
Added necessary config to deploy to skao-rucio-analysis, but haven't deployed yet. Bit concerned about deployment mechanism being a bit too complicated.
Rucio dev meeting in afternoon. Some discussion re: Hermes2.
Migrated dev users over to SDCSS. Few problems along the way, but sorted now. Reset databases and handed over to Philippa.
Started looking at Rucio policy package but got scared.
IRIS meeting.
Made skao-rucio-analysis machine and started adding mechanism to deploy from rucio-analysis.
Bit more investigation of Rucio external access. Spoke to Dimitrios on Slack. He said whoami uses a redirect, so maybe it's that. Upload seems to work.
Updated ska-rucio-client with new rucio/auth hosts.
Updated SDCSS w/ new cubes. About to migrate dev users, but waiting on confirmation no-one's using it.
Got external Rucio analysis to partially work. Some endpoints don't seem to work and others do.
Wrote up some brief notes on probes.
Looked at external access for Rucio.
Looked at external access for rucio. Fought with src-dmz most of day, kept getting logged out / locking terminals.
Looked at rucio probes. Basically a series of python scripts that output to stdout or database, or perform some rucio functionality. Managed via a "cron" replacement jobber (https://dshearer.github.io/jobber/doc/v1.4.
Seems to be random bits of (~zero documented) bespoke functionality w/ hard-coded urls using a mix of diff tech (graphite, prom).
https://github.com/rucio/probes/blob/master/cms/check_unlocked_replicas_per_rse (no comments) https://github.com/rucio/probes/blob/master/common/service_monitoring.py (hard-coded urls) https://github.com/rucio/probes/blob/master/common/check_unlocked_replicas (prom)
Made Rucio PRs for helm chart and container. Merged.
Finished tests of backups. All OK.
Added new volume, cronjob. Investigating some errors with image pull.
Presentation.
AL
Prep for GridPP presentation.
Helped Rohini with switching Rucio servers.
Philippa has spreadsheet of user details. Added bulk add user endpoint to ingest this.
Helped Rohini (try) diagnose NFS issue on rucio development server
Deployed new version of rucio-analysis to escape-rucio-analysis & fixed a few bugs.
Rucio-analysis refactor:
Moved to ansible to manage install and cron, adds possibility of remote deploy.
Removed some code duplication.
Better commented code
Updated README
Rucio-analysis refactor:
Removed job specific shell scripts in favour of docker init script