1. 19 Oct, 2017 1 commit
    • Stevan Radakovic's avatar
      LAVA-876 - Remove access to Dashboard · 103a44be
      Stevan Radakovic authored
      Remove all entries from urls.py in dashboard_app.
      Remove some of the obsolete tests.
      Remove some of the code so that rest of the server works ok.
      
      Change-Id: I74f79f5ce02c4af00416e57299cbe2782fad7ddb
      103a44be
  2. 11 Sep, 2017 1 commit
  3. 17 Jul, 2017 1 commit
  4. 29 Jun, 2017 1 commit
  5. 28 Jun, 2017 1 commit
  6. 08 Jun, 2017 2 commits
  7. 23 May, 2017 1 commit
    • Guillaume Tucker's avatar
      Fix typo which causes XML-RPC error when saving device dict · e23bb176
      Guillaume Tucker authored
      Fix "raise xmlrpclib.Fault(..." in lava_scheduler_app/api.py.  The
      error seen was:
      
        <Fault -32603: "Internal Server Error (contact server administrator for details): 'module' object has no attribute 'Faul'">
      
      This code is only run when the device dictionary fails to be saved on
      the disk, for example due to a permission error if /etc/ is owned by
      root but the server is running as lavaserver.
      
      Fixes: ae4d4776
      
      
      Change-Id: Ic5c480bf610b9e87cdb74502fcd6972c496400aa
      Signed-off-by: Guillaume Tucker's avatarGuillaume Tucker <guillaume.tucker@collabora.com>
      e23bb176
  8. 17 May, 2017 1 commit
    • Neil Williams's avatar
      Expand zmq_client to handle publisher socket · 12c0508b
      Neil Williams authored
      Use the hostname and XML-RPC to lookup the publisher socket
      as the string is not readily visible on most instances.
      Drop authentication requirement for publisher socket as it
      is a copy of a read-only value.
      
      Change-Id: Iad8423e2472196ace62712598e3c13d60ac82d75
      12c0508b
  9. 15 May, 2017 1 commit
  10. 02 May, 2017 1 commit
  11. 27 Apr, 2017 1 commit
    • Florent Jacquet's avatar
      api: add get_recent_jobs_for_device · a04086dd
      Florent Jacquet authored
      
      
      This API function returns the last jobs asked for a given device as a list of
      dictionaries containing the job ID, its description (the name), and its status.
      The typical use case is to build a custom reporting tool outside LAVA by
      querying this function, and then asking for job details once you know which job
      to look for.
      
      Change-Id: I2acd5415ed6ff90d2f64e40d520d79d09c2f4cf3
      Signed-off-by: default avatarFlorent Jacquet <florent.jacquet@free-electrons.com>
      a04086dd
  12. 18 Apr, 2017 1 commit
  13. 14 Feb, 2017 1 commit
    • Rémi Duraffort's avatar
      Generate less database queries · ace40b01
      Rémi Duraffort authored
      In fact, num_devices_visible_to is too complex and generates many queryes for
      most use cases.
      
      Change-Id: I5dd47e0437af5e80cbb7d931ddc72e507ef6d19b
      ace40b01
  14. 24 Jan, 2017 1 commit
  15. 10 Jan, 2017 1 commit
  16. 07 Dec, 2016 1 commit
  17. 01 Dec, 2016 1 commit
  18. 24 Nov, 2016 1 commit
    • Neil Williams's avatar
      LAVA-798 - lookup device-types by alias · dd58a118
      Neil Williams authored
      Allow admins to configure aliases for device-types and provide an API
      to lookup the device-type name by the alias (or a string contained
      within the alias name).
      Add documentation of how this can be used to relate device tree
      names to device-type names whilst keeping device-type names which
      are easily understood by humans. Device types can have any number
      of aliases and some aliases can relate to multiple device types.
      
      Change-Id: I1b2e877acb810addf82c4cafbc214b889a5cd110
      dd58a118
  19. 21 Nov, 2016 1 commit
    • Stevan Radakovic's avatar
      all_devices() api call to do one big query instead of many small ones. · 3e8dbfec
      Stevan Radakovic authored
      device.is_visible_to was called for all devices which generated
      approx. num_of_devices * num_of_devices_of_specific_device_type
      queries before filtering out the RETIRED devices.
      Currently on production that's approx. 100,000 queries per
      all_devices api call.
      This patch also reduces the number of queries to the device_type
      and test_job table for each device.
      
      Change-Id: I9e1b54a088ae6173f337a0e4690484d952d57df7
      3e8dbfec
  20. 08 Nov, 2016 1 commit
  21. 15 Sep, 2016 1 commit
  22. 02 Sep, 2016 1 commit
  23. 29 Aug, 2016 1 commit
  24. 23 Aug, 2016 1 commit
  25. 17 Aug, 2016 1 commit
    • Neil Williams's avatar
      LAVA-200 - track admin actions made in the UI · 2be38d15
      Neil Williams authored
      Create a django admin log record for operations made through the
      UI including restrict and derestrict, put online, put offline,
      put into looping and cancel reserved so as to also log XML-RPC.
      Add explicit logging for updating the device dictionary.
      Note that admin operations in the developer shell cannot be tracked.
      Remove the error in restrict and derestrict which fails due to
      the attempt to create a device state transition from and to the
      same state.
      
      Change-Id: Id5f8efb9078d20fbebfdcbce9919c66940ac58ca
      2be38d15
  26. 05 Aug, 2016 1 commit
    • Neil Williams's avatar
      Django1.10 fixes · b6043723
      Neil Williams authored
      LAVA-441 - drop render_to_response({}, RequestContext(request)
      Switch to loader.get_template(), HttpResponse(template.render({}, request))
      render_to_response with a RequestContext has been removed from
      django1.10 and raises an exception related to the context_processor.
      Drop use of SumIf and replace with conditional expressions.
      Handle tightening of filters on isnull conditions when blank
      is allowed for the foreignkey. (Prevents jobs starting with 1.10)
      
      Change-Id: I59cf9caa0fe3a82027dec0caad5993cda158791b
      b6043723
  27. 13 Jul, 2016 1 commit
    • Neil Williams's avatar
      Allow for creating devices already offline. · da8fa74c
      Neil Williams authored
      Closes: #2419
      Add handling for a missing DeviceStateTransition if a device
      is created already offline. Move the scheduler api unit tests
      to a new file and add a unit test.
      
      Change-Id: I721a15a8345ed41027a50bbc750d391830e7739d
      da8fa74c
  28. 03 Jun, 2016 1 commit
    • Neil Williams's avatar
      Fix error in efa0f5 · 5ca0e29a
      Neil Williams authored
      Keep the pipeline job output consistent with the documented
      interface dictionary.
      Check for a results_link first for V1 jobs, quicker than raising
      and then catching the TypeError exception.
      
      Change-Id: I7a311a5a24230c473f38a908aac9516264278792
      5ca0e29a
  29. 19 May, 2016 1 commit
  30. 11 May, 2016 1 commit
    • Neil Williams's avatar
      Add a bulk job status XML-RPC query · efa0f5ee
      Neil Williams authored
      Querying status of individual jobs is slow, so add a call which
      takes a list and optimise the function to get all the status
      data in one query. Ensure multinode sub_ids are handled cleanly.
      
      Change-Id: I52f0e1c268ad62638ab7d3874271627530cc24ba
      efa0f5ee
  31. 09 May, 2016 1 commit
    • Neil Williams's avatar
      Drop heartbeat support · 97488187
      Neil Williams authored
      Make the worker information more useful for V2 use cases
      whilst retaining essential V1 scheduler tasks and improving
      performance.
      
      Change-Id: I8f365c6b1037b37ea2954da49e5dab64329ffa9e
      97488187
  32. 28 Apr, 2016 1 commit
    • Neil Williams's avatar
      pylint-django changes. · d6859078
      Neil Williams authored
      Identified using pylint-django plugin, only in unstable/stretch.
      
      Change-Id: I65ca9fde297363aed965d5789b8e10692c839ecb
      d6859078
  33. 25 Apr, 2016 1 commit
  34. 20 Apr, 2016 3 commits
    • Quentin Schulz's avatar
      lava_scheduler_app: api: Add get_device_status · 0d7742db
      Quentin Schulz authored
      
      
      Add the get_device_status endpoint to the API to get current status
      (idle, offline, running, ...) of the given device.
      If its status is offline, the date since which it is offline and the
      user responsible of this offlining are joined to this response.
      If its status is running, the job id of the running job is attached to
      this response.
      
      The queried device has to be visible to the user who makes the request.
      
      Change-Id: Ie5c9cba1ebb294e6fe49e7f5d63d7dd1f492b909
      Signed-off-by: default avatarQuentin Schulz <quentin.schulz@free-electrons.com>
      0d7742db
    • Quentin Schulz's avatar
      lava_scheduler_app: api: Add put_into_maintenance_mode · 6da3b4ab
      Quentin Schulz authored
      
      
      Add the put_into_maintenance_mode endpoint to the API to put a given
      device in maintenance mode with the given reason and optionally the
      mail address to optionally notify, if a job is currently running, when
      it has finished.
      
      To access this endpoint, you have to be authenticated as a user with
      sufficient permission (the same as the one used in the web interface:
      superuser, device owner or 'can change device' permission).
      
      The mail address to notify is optional.
      
      Change-Id: I6e5b8e7671f405a56286e0626e1e05e7efe2fbc6
      Signed-off-by: default avatarQuentin Schulz <quentin.schulz@free-electrons.com>
      6da3b4ab
    • Quentin Schulz's avatar
      lava_scheduler_app: api: Add put_into_online_mode · 2091ac9c
      Quentin Schulz authored
      
      
      Add the put_into_online_mode endpoint to the API to put a given device
      in online mode with the given reason and skip health check if asked.
      
      To access this endpoint, you have to be authenticated as a user with
      sufficient permission (the same as the one used in the web interface:
      superuser, device owner or 'can change device' permission).
      
      Skipping health check is optional.
      
      Change-Id: I8b6ffb3b2da00d3ce8940ddb96793726e981a35d
      Signed-off-by: default avatarQuentin Schulz <quentin.schulz@free-electrons.com>
      2091ac9c
  35. 12 Apr, 2016 1 commit
  36. 06 Apr, 2016 1 commit
    • Senthil Kumaran S's avatar
      Fix minor things which affect submit_job xmlrpc api. · 48f41342
      Senthil Kumaran S authored
       * s/JSONDecodeError/simplejson.JSONDecodeError/
       * Fix api doc string to mention YAML where necessary
       * xmlrpclib.Fault should return meaningful content
       * add lava-lxc protocol in YAML job validation schema
      
      Change-Id: I497343fecb347aee83094cb1df610a6808127e2e
      48f41342
  37. 29 Feb, 2016 1 commit
    • Neil Williams's avatar
      Enable pipeline health checks in scheduler · ae26c2a8
      Neil Williams authored
      LAVA-542 - move code out of the scheduler daemon to be
      accessible to dispatcher master and pipeline jobs.
      Adapt health check code to allow for YAML submissions.
      Only change the database job status for a health check,
      the slave dispatcher does not need to know.
      Tidy up unused imports.
      Add support for health check frequency configuration.
      Update health check docs.
      Update status count for health checks.
      
      Change-Id: I4fab34ed571a19062d20936f107866dbf27558be
      ae26c2a8