Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Marius van den Beek
    @mvdbeek
    I guess that depends on whether you can reach your host via docker.for.mac.localhost or host.docker.internal
    I haven't actually tested the PR
    depending on the setup the amqp_url might also need to use that host ?
    John Chilton
    @jmchilton
    I think the problem would become clear if Pulsar logging was fixed in that container
    Maybe I'm wrong though
    Marius van den Beek
    @mvdbeek
    that'd be good either way :)
    John Chilton
    @jmchilton
    Fair :)
    Nate Coraor
    @natefoo
    Anything I should be aware of before updating Main to 19.05?
    @dannon anything I need to do for GIEs?
    Dannon
    @dannon
    Client build should handle it all.
    (so, no, I hope)
    Martin Cech
    @martenson
    🤞
    Nate Coraor
    @natefoo
    make client-production-maps?
    John Chilton
    @jmchilton
    We dropped support for drmaa and Slurm - so you'll need to upgrade your compute nodes to be Kubernetes workers
    Dannon
    @dannon
    client-* yeah.
    Nate Coraor
    @natefoo
    @jmchilton no probs
    @dannon thanks!
    Dannon
    @dannon
    Let me know when it’s deployed and I can kick the tires to make sure it’s all okay.
    Nate Coraor
    @natefoo
    Sounds good
    Nate Coraor
    @natefoo
    Main is on 19.05 now.
    I noticed this in dev too, it looks like jobs are printing the job working dir to stdout?
    Jennifer Hillman-Jackson
    @jennaj
    I've seen this error at main but nate usually "fixes it". Not sure what to tell this user running a local: https://help.galaxyproject.org/t/problem-with-humann2-running-in-galaxy-1-9/1275
    "Conda dependency seemingly installed but failed to build job environment."
    whoa, came out bold. sorry!
    Jennifer Hillman-Jackson
    @jennaj
    Marius van den Beek
    @mvdbeek
    @natefoo galaxyproject/galaxy#7954 should fix the job working dir in stdout issue
    should we trigger another jenkins tool test run again main/test ?
    Nicola Soranzo
    @nsoranzo
    I've extended the timeout for docker-framework builds on Jenkins from 90' to 100', it has aborted a few times recently.
    Nate Coraor
    @natefoo
    @jennaj the slurm packages are me, I'll fix that one.
    @mvdbeek let's give it a bit until I figure out whether I'm going to upgrade again, there was a bad merge somewhere that brought a lot of 19.09 in to the 19.05 branch
    Marius van den Beek
    @mvdbeek
    Screen Shot 2019-05-14 at 14.38.10.png
    that's on usegalaxy.org, seems we still use require.js somewhere
    I also see that locally
    Dannon
    @dannon
    It’s just in the template.
    Oh, wait, locally at what rev? galaxyproject/galaxy#7810 should have removed them.
    Marius van den Beek
    @mvdbeek
    ah, I thought we targeted 19.05 as well
    that was 19.05 local
    Dannon
    @dannon
    I can backport that if we want. They’re harmless, if scary.
    Marius van den Beek
    @mvdbeek
    let's do that, I got confused by this when setting up the IEs
    Dannon
    @dannon
    Ok.
    Nate Coraor
    @natefoo
    What do we want to do about the bad merge to 19.05? @jmchilton suggested just adding all the new stuff to the release notes.
    Dannon
    @dannon
    I think that’s the plan. PR open.
    Nate Coraor
    @natefoo
    Ah ok.
    Dannon
    @dannon
    Not really any need to run tests since it’s just the version string change, but I figured we’d let it sit open for a bit just in case anyone has reservations.
    Martin Cech
    @martenson
    It feels pretty stupid to ransom a repo that everybody has a copy of.
    Dannon
    @dannon
    Right? It seems counterintuitive to the whole design of git.
    Martin Cech
    @martenson
    In cornercases it might work, but I would ransom elsewhere :D