On Mon, May 4, 2009 at 1:35 AM, Tom Haynes <Thomas.Haynes at sun.com> wrote:
> Matty wrote:
>>
>> I configured a new NFS share today with the sharemgr utility, and for
>> some reason the settings are causing the nfs/server SMF service to
>> fault. Here is the configuration I created:
>>
>> $ sharemgr show -vp
>> default nfs=()
>> zfs
>> build nfs=() nfs:sys=(ro="192.168.1.0/24,anon=0")
>> ? ? ? ? ?/bits/provisioning ? ?"Build infrastructure"
>>
>
> I think that share is wrong. You want an '@' in there to denote a subnet:
I accidentally pasted the wrong sharemgr output in the previous
e-mail. The output should have been:
$ sharemgr show -vp
default nfs=()
zfs
build nfs=(anon="0,r...@192.168.1.0/24")
/bits/provisioning "Build infrastructure"
Sorry about that.
> I don't understand this. If we look at the code:
>
> ?557 ? ? ? ? ? ? ? ? ?(void) endnetconfig(nc);
> ?558 ? ? ? ? ? ? ? ? ?if (protoFound == FALSE) {
> ?559 ? ? ? ? ? ? ? ? ? ? ? ? ?fprintf(stderr,
> ?560 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?"couldn't find netconfig entry for
> protocol %s",
> ?561 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?proto);
> ?562 ? ? ? ? ? ? ? ? ?}
>
> then we would expect to see the offending protocol listed in the server log.
> ?'4' is not
> what I would expect...
>
> In your /etc/default/nfs file, is this changed?
This turned out to be the issue. I was experimenting with NFS version
3 and 4, and forgot to put the ALL back in the following item:
NFSD_PROTOCOL=ALL
> BTW: Quick survey, did you look at share_nfs(1M) for the network syntax? If
> you did, was
> it misleading. If you haven't, could you and respond about whether or not the
> syntax worked
>for you?
> I.e., we are concerned that coverage isn't cutting it.
I will take a look through it in the morning and send you my comments.
Thanks for the reply!
- Ryan
--
http://prefetch.net
