Date: Sat, 7 Oct 2023 12:08:14 -0700 From: Rick Macklem <rick.macklem@gmail.com> To: J David <j.david.lists@gmail.com> Cc: FreeBSD FS <freebsd-fs@freebsd.org> Subject: Re: FreeBSD 13.2 NFS client mount hangs Message-ID: <CAM5tNy5H%2B7wZVfVupW2MMt61CQ_tOW--rGrYKn_Ob7u=D1MEXA@mail.gmail.com> In-Reply-To: <CABXB=RS466mdJvPLBZw0MfL%2BgzPjSr8kiHoRiCcdAveNxnGUgg@mail.gmail.com> References: <CABXB=RRSHMhZQFL28eHKjhAYmU87qjpQ=B1=8VRSZoXat9=r5A@mail.gmail.com> <CAM5tNy4sqc18UCZF0vgL%2BXP6vF0wgt_3Yi07yY4wqeuzs6haMA@mail.gmail.com> <CABXB=RSUJ3mpYF5puAm0hSxeavozxyf7Ruab8mPrtBOu6bxM-w@mail.gmail.com> <CAM5tNy52x2s=9Os%2BPAa=-iz7F_o_4_9XxJbRAR28V1v9A4nN6A@mail.gmail.com> <CABXB=RT14gofYHkMMr8cj%2BTy2QRUgn6zunho4T2Kq2NxAWmuAQ@mail.gmail.com> <CAM5tNy7AK7Hq%2Bnnxftg=s5wk=YHw27F7EBydLX=D_z1KvFuD_Q@mail.gmail.com> <CABXB=RS466mdJvPLBZw0MfL%2BgzPjSr8kiHoRiCcdAveNxnGUgg@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Oct 7, 2023 at 10:30=E2=80=AFAM J David <j.david.lists@gmail.com> w=
rote:
>
> Here is a small pcap file obtained during a hang. This hang "feels"
> different from the others. It was a different type of server and the
> processes affected were different. But the symptoms are the same. (As
> are the mount flags, etc., except this fs is mounted read-only.)
>
> It's not much; only 37 packets matched over about five minutes. I'll
> try to get one I'm more confident is of the type previously discussed.
>
> Nonetheless I figured I'd pass this along in case it is helpful and/or
> to confirm that this is the format you're looking for.
Doesn't tell us much. If you look at it in wireshark (I just download
and install it on a windows laptop), you'll see that the very first packet
is the NFS server closing down the TCP connection (a FIN,ACK on a
connection from port 800).
Then the client establishes a new TCP connection and does Renew
operations (just a Sequence operation by itself). Since there is no other
RPCs than the Renews, I cannot guess why the server closed the
connection, nor what the client is up to.
The Sequence reply does note that the callback path is down, which
is expected, given you are not running the nfscbd(8) daemon.
One thing you could try is running the nfscbd(8) daemon.
I suggested against doing so to avoid delegations, but it
might be that this server does not handle that case well.
To do so, add:
nfscbd_enable=3D"YES"
to the NFS client's /etc/rc.conf.
--> It needs to be running before the mount is done, so the
NFS client either needs to be rebooted or the NFS mount
dismounted, the nfscbd started and then the NFS mount
done again.
Maybe this server will be happier with the callback path working?
rick
>
> Thanks!
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAM5tNy5H%2B7wZVfVupW2MMt61CQ_tOW--rGrYKn_Ob7u=D1MEXA>