dinghy: NFS mounting fails when I re-launch dinghy
I have been facing the problem these days. I start dinghy and then stop it using dinghy halt, and then when I start it again, the logs will be like as follows:
Starting the dinghy VM...
Starting NFS daemon, this will require sudo
Waiting for NFS daemon...
NFS mounting failed
Starting the FsEvents daemon
Starting DNS and HTTP proxy
The only solution is to restart my laptop. But once I re-launch dinghy, NFS always fails to mount. Is there anything wrong with my settings and workflow? What should I do to fix it?
- MacOS: 10.13.3
- dinghy: 4.6.3
About this issue
- Original URL
- State: open
- Created 6 years ago
- Reactions: 2
- Comments: 33 (9 by maintainers)
Darn. I’ve tried many times, but I can’t reproduce it myself, so I’ll need some help debugging the issue.
As I understand it, if the UDP port isn’t being freed after the process exits, that means the process is a zombie that hasn’t been fully cleaned up, usually either because it has a child process still running, or because the parent process failed to clean up after it.
Because we daemonize unfsd, the parent process is launchd (PID 1), so cleaning up after it should always work. On the other hand, I don’t think that unfsd ever spawns child processes, so that rules out both possibilities and it seems likely that I’m missing something crucial. 😬
The next time this happens, can you use that netstat command to find the defunct PID, and then look for any information on the parent or children processes? I think
ps -axf | grep <PID>should give relevant info, including parent PID, or you could look in Activity Monitor.@codekitchen I’m using xhyve and getting the same issue as @k0pernicus
Firewall is accepting unfsd
dinghy-NFS.output is telling me
I use
$ netstat -vp udp | grep 19091to get the pid but when I try$ kill 11615the result isHaven’t found a solution yet.
Funny enough: The problem doesn’t show with xhyve. Just tried it, and I’m able to restart dinghy as often as I like…