Closed WIStudent closed 6 years ago
I've not seen this issue on Ubuntu 17 or Fedora 16 or 17 (variously on a Pi, an Intel core i5 and i7). I suspect you have IPv6 disabled in some way (but not completely) at an OS level and it's causing issues. It's a bit of a nightmare in most cases when you try to 'kill' IPv6. There's no specific NFS setting related to the underlying network protocol so there's no anything I believe I can do to help with this. If you disagree, let me know.
Bar telling you to disable IPv6 I think your only option is perhaps to not use --net=host
and map the port specifying only an IPv4 address -p x.x.x.x:2049:2049
. Perhaps also throw in --ipv6=false
.
I'm not convinced that will work so perhaps you could disable IPv6 in the container OS itself by adding this to your run command: --sysctl net.ipv6.conf.all.disable_ipv6=1
. I'll hopefully have a go myself soon.
I'm also experiencing this.
Probably related to this one: https://www.spinics.net/lists/linux-nfs/msg59728.html
Starting rpcbind (uncommenting the code in nfsd.sh
) seems to do the trick
Thank you for your valuable input Willie! I'll get this sorted today and try to remember to check if its still a problem at some point in the future.
OK, rpcbind enabled and latest updated plus tag 7 created.
Hi, I noticed that the startup takes quite a while:
Actually the only thing that needs so much time seems to be the creation of an AF_INET6 TCP socket. On my machine it always takes about 189 seconds. Do you have an idea what the issue could be? It has probably something to do with IPv6 but giving the container an IPv6 address didn't help in my case.