Attending: Will, Petr, Frances, Josh, June, Sebastien, Wilma, Jean-Marie, Simon, Jason, Dominik, David, Chris, Erin, Melissa, Emil, Kevin,
Start: 2:00 pm UK
(2-3 minutes each)
- Next release: 2 Datasets, still at testing stage. idr0095 and maybe idr0093 (might swap)
- Plan for 2-3 weeks time.
- Press release discussions in progress
- Some good progress from JM on notebooks to run segmentation analysis on idr0033.
- Simon & Dom progress on microservices - Image-region ms. Use info from Chris with this and other microservices
- Discussion on HCS data layout for zarr. Converging on 2 good options.
- Hope to decide on best option within a week, and start working on it.
- Josh: in-touch with Trevor re: presentation on Tuesday meeting.
- OMERO 5 / Bio-Formats 6
- JM: release process may change wrt outage. More automated steps.
- David: held-up with ci down
- Josh: move to release candidates if that helps.
- Image-region work ongoing.
- Will check for Glencoe mirrors having various OME jars.
- Seb: looks like latest OMERO releases are there.
- Also use scijava for historical releases
- Global BioImaging workshops in planning - likely early next year
(5 min. max; tech. Discussion should be highlighted to relevant people and rescheduled)
Main Topic: Site-wide outage
- All Dundee services down
- Still waiting to learn full impact.
- Triage affected services - prioritise external public services. Artifactory, website etc.
- Discussing how to restore, migrate etc, dependent on data recovery
- Potentially redirect to mirrors for artifacts etc.
- Might be able to restore, or could need to use tape backups.
- Other lifesci services not affected - except NPSC? - Same hardware?
- Chris: same location.
- Josh: at some point, need to decide whether to move away from UoD services.
- Jason: issues with power supply reliability.
- Simon: waiting to hear from Dell wrt quote and feedback from logs.
- Seb: positive spin: moving to use hosted platforms will be beneficial.
- More robust long-term and could simplify various tasks
- Josh: Glencoe has experience using AWS for stats etc.
- Simon: hook into UoD authentication?
- Chris: not tried it with Azure, but it should ‘just work’
- Jason: do we know how to purchase Microsoft Azure services?
- June: needs to go through UoD purchasing.
- Simon: website can go to github pages
- Josh: we’ll need to run our own artifactory server. Where?
- Simon: can someone else run it for us?
- Chris: remember all URIs must remain identical
- Jason: start to action any external sites? Glencoe becomes public
- Chris: Yes, we have all recent versions.
- But URIs not identical. Different repository names
- Josh: but we can create virtual repository names
- Seb: can make this work for 99% of public artifacts
- Jason: can we start to plan this? Important for community
- Seb: also push artifacts to maven central - redundancy is OK
- Next level of priority is Dundee services - nightshade. Will depend on recovery status.
- Josh: spin up nginx server for redirects?
- Simon: can ask UoD for VM for this. Not needed to do this recently.
- Chris: we can easily set up Digital Ocean VM easily
- Josh: setting up redirects is quite heavy-handed - need to decide what needs doing.
- Jason: sounds like a few actions decided.
(20-25 minutes plus 15 minutes questions max)