Hey everyone! I'm experiencing odd behavior in my ...
# fleet
Hey everyone! I'm experiencing odd behavior in my Fleet deploy and would like to let you know as I think it's a bug: Recently there was a change in the users on my machine (a Ubuntu 22.04): a new user was added to the system (an active user whose shell is
). I expected this change to be reflected on the host details page in Fleet, but it never happened. I refetched a few times and nothing happened, until I decided to remove the machine from Fleet and wait for it to join again. Now that the machine has rejoined, users no longer load. I've already tested the
query locally, used by the server to populate the host_users table, and it returns the desired users in
. I'm currently using version 4.31.0 of Fleet.
I already looked for error logs on the server, but nothing related appeared. I only found lines related to failures in ingesting values referring to disk space:
Copy code
  "component": "http",
  "err": "error in query ingestion",
  "ingestion-err": "ingesting query disk_space_unix: strconv.ParseFloat: parsing \"73,010000000000005.0\": invalid syntax",
  "level": "error",
  "method": "POST",
  "took": "788.528233ms",
  "ts": "2023-05-11T20:09:53.017595007Z",
  "uri": "/api/v1/osquery/distributed/write"
Can you check for any
errors in the logs?
What happens if you run the query as a live query through Fleet?
There are some
errors in the last 3 hours but my machine's
is not one of them
When I run the query live it brings the results as expected
Thanks @Saulo Guilhermino. Reaching out to the team for some input on this one.
No problem! Thanks for the support
One thing that I should probably have you check (it's always the "obvious" things I miss)... Have you disabled host users, or overridden the detail query?
I haven't really messed with any settings in the meantime
I just removed two more hosts to perform the same test. An Ubuntu with OSQuery 5.8.2 and a Windows with OSQuery 5.8.1. Both have already rejoined but only Windows users are showing (along with disk information)
Excellent test, thank you.
Can you grab the osquery logs from one of those Linux Hosts for me?
Here are the logs for the osqueryd on my machine via journalctl
And here are the logs for the last two days for my host on Fleet
Just wanna chime in and say that I have this issue too on 4.31.0, no changes made to the default host config. Fetching via live query returns expected results, I’ll check the other troubleshooting steps suggested on Monday :)
Sorry @Saulo Guilhermino, I somehow missed those logs!
No problem, Kathy! 😄