Hi James! This sounds like a Docker issue. Where a...
# fleet
m
Hi James! This sounds like a Docker issue. Where are you seeing these error messages?
j
@Mason Buettner in the logs....
We're running fleet under kubernetes with:
Copy code
securityContext:
            allowPrivilegeEscalation: false
            capabilities:
              drop: [ALL]
            privileged: false
            readOnlyRootFilesystem: true
            runAsGroup: 3333
            runAsUser: 3333
            runAsNonRoot: true
The error message suggests that in the new fleet release, something expects a home dir to exist for the current user; and the docker image you guys create doesn't have a homedir and/or a correct /etc/passwd set up
m
Thanks! What version were you updating from? Can you share more info about your infrastructure?
j
Updating from the previous version (i.e.
v4.83.2
)
What else would you like to know? I don't think much would be relevant
m
Broad strokes are fine, I'm asking so we can try to replicate the behavior.
j
I expect you can replicate by running the official fleet docker container with the uid and gid set to 3333
👀 1
Copy code
$ sudo docker run -it --user 3333:3333 fleetdm/fleet:v4.84.0@sha256:51b56ad59a840b28e074ff9b06d6d5b232b0ca2f0d999bb164820da69c7cbe15
Failed to fetch user info for home directory: user: unknown userid 33332026/04/27 23:12:33 71 <nil>
^ replicated with a plain docker command
Compare to previous release (which obviously starts failing due to not setting db credentials; but does show that we don't get the same home directory error):
Copy code
$ sudo docker run -it --user 3333:3333 fleetdm/fleet:v4.83.2@sha256:4b7746ca86824896503993d97cda9cca952f9c86f268686236858426084187c9
ts=2026-04-27T23:13:47Z level=warn msg="could not connect to db" err="dial tcp [::1]:3306: connect: connection refused" sleep_interval=0s
ts=2026-04-27T23:13:47Z level=warn msg="could not connect to db" err="dial tcp [::1]:3306: connect: connection refused" sleep_interval=1s
ts=2026-04-27T23:13:48Z level=warn msg="could not connect to db" err="dial tcp [::1]:3306: connect: connection refused" sleep_interval=2s
m
Thanks James, I’ll do some research and circle back tomorrow with an update.
j
@Jorge Falcon I just saw your email; I replied and linked this thread.
m
@James just following up on this, it's been determined a bug and a report has been filed here: https://github.com/fleetdm/fleet/issues/44298
❤️ 2
z
hey @James! how are things running since the upgrade to 4.84.1?