Yet another node offline after update thread

Node Operators troubleshooting
, , ,
27

Not on current system, but I did on older servers with no success. I also can’t believe that :slight_smile:
I think my docker setup around storagenode triggers the problem, I am disabling the containers trying to see when the problem will be gone, but it will take months to get the results.
What I mean by that, I have these containers for each node:

  • storjlabs/storagenode (logs to file node.log)
  • anclrii/storj-exporter (linked to storagenode container)
  • storj-log-exporter (grok exporter) (bound to storagenode path and node.log)
  • linkyard/logrotate (bound to storagenode path and node.log)

Maybe some scenario when storj-exporter pulls the data and update happens to stop storagenode at the same time?

28

Unlikely. But definitely some race condition, rare and hard to reproduce.
However I did not figure out why the docker daemon does not restart the stopped container in this case.