Longhorn has high requirement on network latency and bandwidth. The unstable underlying network will result in Longhorn volume failure and it would be hard for users to know it's the underlying network issue or Longhorn storage issue.
As a first step, the tool should be separate from Longhorn, and can be used by the users to determine if the underlying network caused volume failure or not.
@dnauck @botzill @phal0r
I'm running longhorn over an S3 bucket, with the underlying path handled by s3ql
It launches and installs, creates volumes, creates PVs, but never attaches

Update:
Looks like something is unsupported, and not latency, per se
Error: file extent is unsupported
time="2019-12-02T22:06:44Z" level=warning msg="test1gb-r-68fb05d0: time=\"2019-12-02T21:51:36Z\" level=info msg=\"Creating volume /host/s3ql/mnt/rancher/longhorn/replicas/test1gb-eb6a4576, size 1073741824/512\""
time="2019-12-02T22:06:44Z" level=warning msg="test1gb-r-68fb05d0: time=\"2019-12-02T21:51:37Z\" level=fatal msg=\"Error running start replica command: file extent is unsupported: operation not supported\""
time="2019-12-02T22:06:44Z" level=warning msg="Instance test1gb-r-68fb05d0 is state error, error message: exit status 1"
time="2019-12-02T22:06:44Z" level=warning msg="Instance test1gb-r-68fb05d0 crashed on Instance Manager instance-manager-r-4f9245b7 at dream, try to get log"
Answered here, https://github.com/longhorn/longhorn/issues/533#issuecomment-531594365
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
Most helpful comment
@dnauck @botzill @phal0r