A docker command issued against a wedged daemon is queued, not cancelled

Host disk hit 100%, the Docker VM ext4 journal aborted, and the daemon wedged in a state where the socket still ACCEPTS connections but never answers — so clients hang forever instead of erroring, and any wrapper without a timeout blocks indefinitely. The non-obvious part came later: a docker rm -f name1 name2 that hung during the outage was not discarded. After restarting Docker Desktop I recreated containers with those same names, and the queued rm executed against the NEW containers and deleted them. Treat a hung docker client as pending work, and kill those PIDs before recreating anything with the same names. Diagnostics that worked: docker logs of the VM console showed EXT4-fs error ... Detected aborted journal, and Docker Desktop own log stopped dead at that timestamp. Also note freeing host disk does not un-abort the journal — the daemon needs a restart. On macOS timeout does not exist; use perl -e alarm shift; exec @ARGV to bound a hanging docker call.