Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    dpantke
    @dpantke
    That's not in the new version that I updated to.
    mcmiv413
    @mcmiv413
    @anonymuse Thanks :-)
    Yami
    @yamikuronue
    @mccrackend yeah that's what we do under Cattle today, but it's a taks and a half to get the devs to build them
    Jesse White
    @anonymuse
    That's what I would recommend as well @mccrackend . I see people creating "ready" and "live" functionality into their health endpoints and then using their applications to manage when a microservice is up and running (but not able to accept traffic, so 5xx of some sort), and then up with dependencies met and ready to serve traffic (gives a 200).
    @yamikuronue -- good health checks are so important but yes so hard to get right!
    especially in microservices land. it's an art really.
    Dan Mikita
    @danmikita
    @yamikuronue Not sure what stack your using, but there are usually frameworks that you can just add to your dependency tree. Such as Actuator in Spring.
    Yami
    @yamikuronue
    I literally provide a Hapi plugin that handles healthchecks based on a config for them, and they still aren't using it lol
    Dan McCracken
    @mccrackend
    @yamikuronue Am dev. Can confirm we don't like doing work XD jk
    Matt Chesler
    @mchesler
    YAML in, JSON out? -5 points for consistency...
    doceallaigh
    @doceallaigh
    So could you manually edit the previous spec and trick docker into updating to a new spec on rollback?
    Dan McCracken
    @mccrackend
    Need to make devs somewhat accountable for the ops. Then they'll write better healthchecks.
    Jesse White
    @anonymuse
    @doceallaigh -- where would you edit the spec?
    @mccrackend -- that's the idea!
    Yami
    @yamikuronue
    yeah... still working on that part :D @mccrackend
    we're six months in to trying to swallow agile + devops all in one go. It's frustrating but we're making progress
    Steve Huff
    @hakamadare
    @anonymuse a bit of a bug in the slides: if we’ve been doing everything as directed, we previously ran export TAG=v0.2
    in this case Bret expects us to not have $TAG set and so it should default to latest
    Dan Mikita
    @danmikita
    @anonymuse Git repo for all Stack yaml --> 1 branch per environment --> deamon running to check for new commits --> new commit executes the docker command for deploying the new file(s)
    Yami
    @yamikuronue
    code it live, that always works!
    Jesse White
    @anonymuse
    Thanks @hakamadare -- I'll collect that feedback to fix in the slides.
    Steve Huff
    @hakamadare
    :thumbsup:
    the other bug is that the character is named “MacGuyver”, not “McGuyver"
    Dan McCracken
    @mccrackend
    @yamikuronue Agile + DevOps at the same time?? I suppose if mgmt & leadership fully support the flailing that is probably happening internally, might have a shot
    mcmiv413
    @mcmiv413
    @hakamadare Why not include his first name? :-D
    Steve Huff
    @hakamadare
    :trollface:
    Yami
    @yamikuronue
    @mccrackend to be fair, we've been half-agile for years: start eveyr project waterfall, and when the deadline goes flying past due to scope creep, start doing agile XD
    Christian Dahlhausen
    @cdahlhausen
    Looks like docker node ps -a is not an option anymore
    -a
    Dan McCracken
    @mccrackend
    @yamikuronue That actually sounds better than my current "half-ass-all-agile-but-not-really-agile" organization :smile:
    dpantke
    @dpantke
    You mean like every other enterprise @mccrackend ?
    Yami
    @yamikuronue
    I'm sure we're hiring ;P
    Dan McCracken
    @mccrackend
    It might be a dumpster fire, but its MY dumpster fire dammit
    Yami
    @yamikuronue
    amen to that
    Dan McCracken
    @mccrackend
    @dpantke amen brother
    dpantke
    @dpantke
    You always like your own brand. :)
    Any way of partitioning the Raft DB such that I can have multiple service envs running on the same cluster?
    Jesse White
    @anonymuse
    @cdahlhausen -- did you mean docker container ps -a?
    Node is for looking at the available workers for the swarm.
    docker node ls
    Docker node ps looks for the tasks of the cluster, though I can't remember at the moment if it ever had an -a flag.
    Yami
    @yamikuronue
    Slide 194 appears to be wrong
    it's listing docker node ps -a and there's no such flag it seems
    Jesse White
    @anonymuse
    @yamikuronue -- that's a bug yes, we'll fix that.

    I find it useful to look at a set of tasks via the service paradigm

    docker service ps <service_name>

    Jesse White
    @anonymuse
    How's everyone doing?
    Christian Dahlhausen
    @cdahlhausen
    @anonymuse It was referenced on some slide
    Jesse White
    @anonymuse
    @cdahlhausen -- yes, that's a bug now that the API has changed. Not sure when/if it was changed.
    Steve Huff
    @hakamadare
    @anonymuse so far so good!
    Matt Chesler
    @mchesler
    @anonymuse :+1: