Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    EHJ-52n
    @EHJ-52n
    @JohnPTobe Thank you very much for your clarifying response on jul 26. Have a nice day.
    Juozas Gaigalas
    @juozasg
    how difficult is to upgrade to 5.4.0 to 5.9.5?
    John Tobe
    @JohnPTobe
    Not terribly difficult, but long. It can take several hours to run the migrations. You'll want to backup your database first and give yourself a window where it won't impact production too much.
    Juozas Gaigalas
    @juozasg
    got it. can we do it by changing configs in marathon? is there an example of how to initiate the upgrade?
    John Tobe
    @JohnPTobe
    Pause the scheduler; let jobs finish running. This is so messages don’t get dropped
    Back up the db and stop scale services
    Scale, the logger, scale webserver... I think that’s it
    Juozas Gaigalas
    @juozasg
    is this part that needs to be updated in marathon to scale:5.9.6? other configs as well?
    {
      "type": "DOCKER",
      "docker": {
        "forcePullImage": true,
        "image": "geoint/scale:5.4.0",
        "parameters": [],
        "privileged": false
      },
      "volumes": []
    }
    John Tobe
    @JohnPTobe
    Once the services are stopped and the db is backed up then yes, change that to 5.9.5 and start the scale service again
    Juozas Gaigalas
    @juozasg
    stuff like this doesn't change:
    LOGSTASH_DOCKER_IMAGE=geoint/scale-logstash:5.4.0
    John Tobe
    @JohnPTobe
    It should start the scale log service and webserver
    Oops, change that too
    Juozas Gaigalas
    @juozasg
    to 5.9.5 or 5.9.6, right?
    John Tobe
    @JohnPTobe
    Right, I don’t think there is a 5.9.6
    It should do the migrations and run other db updates
    @gisjedi does that cover it? How long did it take back in March?
    If something goes wrong you should be able to restore the db and change back to 5.4
    Juozas Gaigalas
    @juozasg
    it's a demo system so we are not worried about losing data
    John Tobe
    @JohnPTobe
    Cowboy up then 🤠
    Juozas Gaigalas
    @juozasg
    DCOS_PACKAGE_VERSION=5.4.0-0.1.0
    leave as is?
    ```
    John Tobe
    @JohnPTobe
    That I don’t know
    Juozas Gaigalas
    @juozasg
    "RABBITMQ_DOCKER_IMAGE": "rabbitmq:3.6-management",
    John Tobe
    @JohnPTobe
    I think that should stay the same
    Juozas Gaigalas
    @juozasg
    Failed to launch container: Failed to run 'docker -H unix:///var/run/docker.sock pull geoint/scale:5.9.5': exited with status 1; stderr='manifest for docker.io/geoint/scale:5.9.5 not found '
    use 5.9.4 instead?
    John Tobe
    @JohnPTobe
    Hmm, maybe we never pushed images to docker hub? I don’t know if 5.9.4 is there either then
    Juozas Gaigalas
    @juozasg
    i'm seeing 5.9.4 there
    not 5.9.5
    John Tobe
    @JohnPTobe
    Hmm ok; the upgrade to 5.9.5 is important but easy once you’re at 5.9.4
    So I can see about getting 5.9.5 up there
    Juozas Gaigalas
    @juozasg
    ty :)
    Juozas Gaigalas
    @juozasg
    no luck upgrading to 5.9.4
    I0905 12:43:16.284466  9997 executor.cpp:186] Starting task scale.41fcb7c3-cffc-11e9-9c1f-decc67049e72
    /usr/lib/python2.7/site-packages/urllib3/connectionpool.py:857: InsecureRequestWarning: Unverified HTTPS request is being made. Adding certificate verification is strongly advised. See: https://urllib3.readthedocs.io/en/latest/advanced-usage.html#ssl-warnings
      InsecureRequestWarning)
    No handlers could be found for logger "marathon"
    Traceback (most recent call last):
      File "bootstrap.py", line 380, in <module>
        run(client)
      File "bootstrap.py", line 60, in run
        deploy_rabbitmq(client, rabbitmq_app_name)
      File "bootstrap.py", line 319, in deploy_rabbitmq
        if not check_app_exists(client, app_name):
      File "bootstrap.py", line 163, in check_app_exists
        client.get_app(get_group_app_name(app_name))
      File "/usr/lib/python2.7/site-packages/marathon/client.py", line 253, in get_app
        'GET', '/v2/apps/{app_id}'.format(app_id=app_id), params=params)
      File "/usr/lib/python2.7/site-packages/marathon/client.py", line 113, in _do_request
        raise MarathonHttpError(response)
    marathon.exceptions.MarathonHttpError: MarathonHttpError: HTTP 401 returned with message, "Unauthorized"
    I0905 12:43:17.389161 10000 executor.cpp:736] Container exited with status 1
    updated scale-logstash service config to 5.9.4
    Jonathan Meyer
    @gisjedi
    @juozasg Can you try the configuration available here and see if it helps? https://github.com/ngageoint/scale/wiki/Deploy-v5-Recommendations
    Sorry for the missing images. Our primary environment does not consume the Docker Hub images, so build issues are sometimes missed there. We've been in heavy development for 7.x releases and missed the errors in the 5.9.x Docker Hub builds. I think I have a fix in for that... waiting on the build to complete and confirm.
    alaisslx
    @alaisslx
    @gisjedi @JohnPTobe thank you very much for the support and updated documentation.
    John Tobe
    @JohnPTobe
    @juozasg @alaisslx FYI the build has been fixed and 5.9.7 is now available
    Juozas Gaigalas
    @juozasg
    i was able to install scale 5.9.7 using those directions. seeing some minor issues that i think can be fixed by rebooting a node. thank you!
    John Tobe
    @JohnPTobe
    🎉
    JimHerner
    @JimHerner
    Howdy all, I've been fidgeting with a Scale 5.9.7 deployment for a bit now and seem to be having some trouble getting recipes with PARSE triggers to run. We can get recipes with INGEST triggers to run fine, and it seems like our parse-job is correctly tagging the source files with the data_types that the PARSE trigger is searching for, but the recipe with the PARSE trigger never runs, and the source files appear to stay in the ingesting folder in the workspace (as opposed to being deferred like we've seen for the INGESTed files that don't trigger any existing trigger rules). Any thoughts?
    Jonathan Meyer
    @gisjedi
    Jim, sorry to hear your running into trouble. What you are doing sounds reasonable. Have you looked through the scheduler logs to see if there are any errors related to the firing of the trigger? Feel free to DM me the stdout/stderr from the scale service in your cluster and I can see if anything jumps out at me.
    ants-chavez
    @ants-chavez
    Scale team, I have been having an issue the last couple of weeks where the the Scale main page shows 0 active nodes, and when you click on the nodes page it just hangs and never comes back.
    Jobs are actively processing and everything else seems to be working.
    Is there a service I can restart or something I can do to fix this without having to redeploy Scale from DC/OS?
    I am currently using version 5.5.0
    emimaesmith
    @emimaesmith
    @ants-chavez I apologize for the delay in response. Sorry to hear you're running into this problem. Do you have any logs (webserver, scheduler, browser console) you could share that might help us pinpoint the source of your issue?
    aaward-professional
    @aaward-professional
    I'm trying to stand up a scale cluster on top of minidcos. I've got all services up and healthy, but trying to access the scale-ui (http:/<my-ip>/service/scale) leads me to a blank page. When I look at the source, it's loading the html, but failing to find any of the scripts the ui depends on - its links to http://<my-ip/pollyfill.js, but that script is accessible at http://<my-ip>/service/scale/polyfill.js.
    Any ideas as to what's going or what to do about it?