redpanda startup issue: docker/swarm- diagnosed/resolved

[thefnordling]

i had been getting a seastar startup error trying to run redpanda via swarm; and had spent a bit of time trying (and failing) to identify the root cause. I had thought the issue was storage related (redpanda worked on numerous other hosts/configs and this vm had a different storage backend). The error i recvd on startup is below:

Could not initialize seastar: std::__1::system_error (error system:1, Operation not permitted)

rpk redpanda check showed no fatal errors, and i couldn't find any logging from seastar as to what the specific cause of the system error was.

Eventually i noticed the hosts were running an out of date version of docker (i had assumed the VMs were newer than they were) and after upgrading docker-ce from 20.10.8 to 20.10.21 everything started up fine.

more expressive seastar logging output or an rpk redpanda check that tests for whatever the underlying fatal error was, or even a check for the a minimum-required docker version would have helped expedite resolution.