Forcing an update failure

#1

Hello,

I noticed a similar post here, but mine is a little different.

I’m using the hosted mender service with a Pi. I would like to demo a rollback in the event that a failed deployment to a Pi device. Is there something I can do to a new deployment to force it to fail and roll back to the original partition? Maybe a one-line config change or deleting a file from an image that is required. Any suggestions would be appreciated.

#2

There are a couple of ways of doing this and what I typically do is to create an Mender Artifact that has bogus payload data.

E.g

dd if=/dev/zero of=tmp.data bs=1M count=1
mender-artifact write rootfs-image -f tmp.data --device-type raspberrypi3 --artifact-name "bad-artifact"
#3

Thanks much, that update did the trick! I was able to force a failure to demo Mender’s rollback.

One follow up question. I’m on a Mac and getting an error when trying to use the mender-artifact to read out a file: mender-artifact cat test.mender:/etc/hosts
failed to copy from: test.mender:/etc/hosts to stdout: err: extFile: ReadError: debugfsCopyFile failed: debugfs: run debugfs dump: exec: “debugfs”: executable file not found in $PATH

Any thoughts?

Thanks again!