Uploaded image for project: 'Subscription Watch'
  1. Subscription Watch
  2. SWATCH-1650

Hosts from Discovery report are not visible in SWatch since early August

XMLWordPrintable

    • False
    • Hide

      None

      Show
      None
    • False

      Hosts from Discovery report are not visible in SWatch.

      Report is generated by Discovery and uploaded through Yuptoo. Hosts correctly appear in HBI. After issuing SWatch sync, they are not available in SWatch API response. I can't tell if they never make it to SWatch, or they can't be mapped correctly to architecture and are filtered out while making an API response.

      We have 2 reports in iqe-foreman-rh-cloud-plugin:

      • 1.0 last passed in August 4 and first failed in August 8
      • 1.1 last passed in August 2 and first failed in August 4

      I don't know why there's a time difference here. I would expect both to pass/fail at the same time, as environment is otherwise unchanged.

      In SWatch tally logs, when testing with report that has single host, I found:

      2023-08-30 09:57:07,818 [thread=rhsm-subscriptions-task-processor-0-C-1] [INFO ] [org.candlepin.subscriptions.task.queue.kafka.KafkaTaskProcessor] - Message received from kafka (key=16781032): JsonTaskMessage(groupId=platform.rhsm-subscriptions.tasks, type=UPDATE_SNAPSHOTS, args={orgs=[16781032]})
      2023-08-30 09:57:07,819 [thread=rhsm-subscriptions-task-processor-0-C-1] [INFO ] [org.candlepin.subscriptions.tally.tasks.UpdateOrgSnapshotsTask] - Updating snapshots for org 16781032.
      2023-08-30 09:57:07,819 [thread=rhsm-subscriptions-task-processor-0-C-1] [INFO ] [org.candlepin.subscriptions.tally.TallySnapshotController] - Producing snapshots for Org ID 16781032 
      2023-08-30 09:57:07,885 [thread=rhsm-subscriptions-task-processor-0-C-1] [INFO ] [org.candlepin.subscriptions.tally.InventoryAccountUsageCollector] - Reconciled 1 records for orgId=16781032
      2023-08-30 09:57:07,909 [thread=rhsm-subscriptions-task-processor-0-C-1] [INFO ] [org.candlepin.subscriptions.tally.InventoryAccountUsageCollector] - Running tally via DB for orgId=16781032
      2023-08-30 09:57:07,996 [thread=rhsm-subscriptions-task-processor-0-C-1] [INFO ] [org.candlepin.subscriptions.tally.SnapshotSummaryProducer] - Produced 0 TallySummary messages
      2023-08-30 09:57:07,996 [thread=rhsm-subscriptions-task-processor-0-C-1] [INFO ] [org.candlepin.subscriptions.tally.MaxSeenSnapshotStrategy] - Finished producing snapshots for orgId=16781032
      

      So there's 1 record reconciled, but no TallySummary messages produced. Which suggests a host is filtered out between these two steps.

      Discovery was supposed to switch from facts.rh_products_installed to system_profile.installed_products, but that has not happened yet. I think it's part of DISCOVERY-226 . Even if Discovery makes a switch, there's a question of customers using older Discovery versions and reports generated by older versions.

            kflahert@redhat.com Kevin Flaherty
            mzalewsk@redhat.com Mirek Długosz
            Mirek Długosz Mirek Długosz
            Votes:
            0 Vote for this issue
            Watchers:
            9 Start watching this issue

              Created:
              Updated:
              Resolved: