0

Canary Historian fatal errors

  I had a couple of problems. One was that my datasets failed the automatic verification, so I had to manually verify them. Annoying, yes, but oh well, not a big deal. Second was that suddenly the Canary Historian started crashing intermittently (yes, the historian service), after running fairly well for a month. I had recently added another 5000 tags to the mqtt collector. So that's the context. I tracked down a series of errors preceeding the historian crash, and found the first one. One of the sparkplug B metrics, a very long string, triggered a Receiver error "80040414 - This trend vector is full". It may have been a concidence, but once I removed these metrics from the Sparkplug payloads, both of my problems were fixed. No more historian fatal errors, and all the datasets were automatically verified. Runs smooth as silk now, and has been since nov 10th. So the problem is fixed now, but I thought I'd let you know that there MAY be a problem related to very large string metrics from sparkplug B. That could cause historian crashing. 

1reply Oldest first
  • Oldest first
  • Newest first
  • Active threads
  • Popular
  • Great catch and thank you Jon for the diagnostic work you put into it. Will get Josh Wolf from our team to take a look.

    Like
Like Follow
  • Status Acknowledged
  • 10 mths agoLast active
  • 1Replies
  • 42Views
  • 3 Following