Re: RFC: NFS over TLS stats

Thanks for all the suggestions. I've modified the patch for all of them.
(tx/rx instead of snd/rcv, _tls at the end of the names so the order is
 more readable.)

There aren't really errors visible at the RPC/TLS boundary, but I have
added counts of successful/failed handshakes and a count of alert
messages (basically "Peer Resets" from what I've seen).

Thanks, rick
ps: Just replying to one of the messages and, yes, I did top post..


On Wed, Oct 25, 2023 at 8:47 AM Freddie Cash <fjwcash@gmail.com> wrote:
>
> On Tue, Oct 24, 2023 at 10:56 PM Rick Macklem <rick.macklem@gmail.com> wrote:
>>
>> Garrett Wollman asked me via email how a server
>> admin could tell what usage NFS over TLS was
>> happening.
>>
>> I admitted that there was nothing. I have come up
>> with a patch that generates the following:
>> kern.rpctls.snd_tls_msgbytes: 21508
>> kern.rpctls.snd_msgbytes: 20828
>> kern.rpctls.snd_tls_msgcnt: 57
>> kern.rpctls.snd_msgcnt: 58
>> kern.rpctls.rcv_tls_msgbytes: 12336
>> kern.rpctls.rcv_msgbytes: 12072
>> kern.rpctls.rcv_tls_msgcnt: 57
>> kern.rpctls.rcv_msgcnt: 58
>
>
> Have no technical comments, but a style suggestion:  put the tls at the end to make the output look nicer.  :)
>
>> kern.rpctls.snd_msgbytes: 21508
>> kern.rpctls.snd_msgbytes_tls: 20828
>> kern.rpctls.snd_msgcnt: 58
>> kern.rpctls.snd_msgcnt_tls: 57
>> kern.rpctls.rcv_msgbytes: 12072
>> kern.rpctls.rcv_tmsgbytes_tls: 12336
>> kern.rpctls.rcv_msgcnt: 58
>>
>> kern.rpctls.rcv_msgcnt_tls: 57
>
>
> --
> Freddie Cash
> fjwcash@gmail.com