|
Replies:
14
-
Last Post:
Mar 22, 2006 5:58 AM
by: ceri
|
|
|
Posts:
192
From:
Registered:
2/21/06
|
|
|
|
Problem accessing snapshot dirs over NFSv4 w/ FreeBSD
client
Posted:
Mar 21, 2006 2:42 PM
|
|
Hi,
After having gotten the other issue sorted, now I'm running into a new one!
I made a snapshot of a user's home directory. I can browse to it just fine
on the Solaris machine running zfs.
# pwd
/export/zfs/chuangj/.zfs/snapshot/03212006
#
# zfs list
NAME USED AVAIL REFER MOUNTPOINT
users 18.5G 166G 98.5K /users
users/home 18.5G 166G 99.5K /export/zfs
users/home/chuangj 3.17G 166G 3.17G /export/zfs/chuangj
users/home/chuangj@03212006 498K - 3.17G -
#
However, on the FreeBSD machine, it's not working.
Filesystem Size Used Avail Capacity Mounted on
server2:export/zfs/chuangj 169G 3.2G 166G 2%
/usr/home/chuangj
%pwd
/usr/home/chuangj/.zfs/snapshot
%ls
03212006
%cd 03212006
03212006: Too many levels of remote in path.
%
%ls -als
ls: 03212006: Too many levels of remote in path
total 0
0 dr-xr-xr-x 3 root wheel 3 Mar 21 12:41 .
0 dr-xr-xr-x 3 root wheel 3 Mar 21 12:41 ..
%
Reading up some, I've seen some references to not being able to mount the
filesystem because it's already mounted. However, this didn't make clear
what I'm doing wrong. :) Once again, I ask for your kind assistance!
Cheers,
David
PS - I'm really loving ZFS so far, this is FUN!
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
Posts:
370
From:
US
Registered:
6/13/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 3:44 PM
in response to: ormandj
|
|
David J. Orman wrote:
>Hi,
>
>After having gotten the other issue sorted, now I'm running into a new one!
>
>I made a snapshot of a user's home directory. I can browse to it just fine
>on the Solaris machine running zfs.
>
># pwd
>/export/zfs/chuangj/.zfs/snapshot/03212006
>#
>
># zfs list
>NAME USED AVAIL REFER MOUNTPOINT
>users 18.5G 166G 98.5K /users
>users/home 18.5G 166G 99.5K /export/zfs
>users/home/chuangj 3.17G 166G 3.17G /export/zfs/chuangj
>users/home/chuangj@03212006 498K - 3.17G -
>#
>
>However, on the FreeBSD machine, it's not working.
>
>Filesystem Size Used Avail Capacity Mounted on
>server2:export/zfs/chuangj 169G 3.2G 166G 2%
>/usr/home/chuangj
>
>%pwd
>/usr/home/chuangj/.zfs/snapshot
>%ls
>03212006
>%cd 03212006
>03212006: Too many levels of remote in path.
>%
>%ls -als
>ls: 03212006: Too many levels of remote in path
>total 0
>0 dr-xr-xr-x 3 root wheel 3 Mar 21 12:41 .
>0 dr-xr-xr-x 3 root wheel 3 Mar 21 12:41 ..
>%
>
>Reading up some, I've seen some references to not being able to mount the
>filesystem because it's already mounted. However, this didn't make clear
>what I'm doing wrong. :) Once again, I ask for your kind assistance!
>
>Cheers,
>David
>
>
>
I suspect that the freeBSD client is using NFSv2 or NFSv3, and that your
server doesn't have support for NFSv3 .zfs/snapshot access yet.
which build is your server? and what version of NFS is your BSD client
using to talk to the server?
If you can grab a snoop trace that would be ideal (on the server as
root, #snoop -o /tmp/snoop <clientname> ).
oh yeah, out of curiosity, what shell are you using on your client?
>PS - I'm really loving ZFS so far, this is FUN!
>
>
>
excellent, keep having fun....
eric
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
192
From:
Registered:
2/21/06
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 3:55 PM
in response to: goo
|
|
> I suspect that the freeBSD client is using NFSv2 or NFSv3, and that your
> server doesn't have support for NFSv3 .zfs/snapshot access yet.
>
> which build is your server? and what version of NFS is your BSD client
> using to talk to the server?
Server:
# uname -a
SunOS server2 5.11 snv_33 i86pc i386 i86pc
#
Client:
%uname -a
FreeBSD server1.corenode.com 6.0-RELEASE-p2 FreeBSD 6.0-RELEASE-p2 #2: Wed
Jan 11 13:52:14 HST 2006
root at server1 dot corenode dot com:/usr/obj/usr/src/sys/SERVER1 i386
%
I think FreeBSD uses NFSv4 by default. How would I check? I just did
"mount server2:/export/zfs/chuangj /home/chaungj" to mount (and added an
fstab entry.)
>
> If you can grab a snoop trace that would be ideal (on the server as
> root, #snoop -o /tmp/snoop <clientname> ).
http://corenode.com/~ormandj/files/snoop.gz
I ran that command, then I tried to cd 03212006, and got the error. I'm
guessing that's what you wanted me to do. :)
>
> oh yeah, out of curiosity, what shell are you using on your client?
%echo $SHELL
/bin/tcsh
%
>>PS - I'm really loving ZFS so far, this is FUN!
>>
>>
>>
> excellent, keep having fun....
>
> eric
>
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
466
From:
US
Registered:
12/21/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 3:59 PM
in response to: ormandj
|
|
David J. Orman wrote On 03/21/06 16:42,:
> Hi,
>
> After having gotten the other issue sorted, now I'm running into a new one!
>
> I made a snapshot of a user's home directory. I can browse to it just fine
> on the Solaris machine running zfs.
>
> # pwd
> /export/zfs/chuangj/.zfs/snapshot/03212006
> #
>
> # zfs list
> NAME USED AVAIL REFER MOUNTPOINT
> users 18.5G 166G 98.5K /users
> users/home 18.5G 166G 99.5K /export/zfs
> users/home/chuangj 3.17G 166G 3.17G /export/zfs/chuangj
> users/home/chuangj@03212006 498K - 3.17G -
> #
>
> However, on the FreeBSD machine, it's not working.
>
> Filesystem Size Used Avail Capacity Mounted on
> server2:export/zfs/chuangj 169G 3.2G 166G 2%
> /usr/home/chuangj
>
> %pwd
> /usr/home/chuangj/.zfs/snapshot
> %ls
> 03212006
> %cd 03212006
> 03212006: Too many levels of remote in path.
> %
> %ls -als
> ls: 03212006: Too many levels of remote in path
> total 0
> 0 dr-xr-xr-x 3 root wheel 3 Mar 21 12:41 .
> 0 dr-xr-xr-x 3 root wheel 3 Mar 21 12:41 ..
> %
>
> Reading up some, I've seen some references to not being able to mount the
> filesystem because it's already mounted. However, this didn't make clear
> what I'm doing wrong. :) Once again, I ask for your kind assistance!
>
> Cheers,
> David
>
> PS - I'm really loving ZFS so far, this is FUN!
>
> _______________________________________________
> nfs-discuss mailing list
> nfs-discuss at opensolaris dot org
Just guessing, I'd say the client is coming across over NFSv3.
And this is a known bug:
6344186 NFSv3 needs to support .zfs (like NFSv4 already does)
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6344186
I think you just need to wait for the changes to make a public
release.
Watch this example:
NFSv3 fails:
[tdh@adept .zfs]> cd snapshot/
[tdh@adept snapshot]> ls -al
total 0
dr-xr-xr-x+ 4 root root 4 Mar 21 17:52 .
dr-xr-xr-x+ 3 root root 3 Mar 21 17:52 ..
?--------- ? ? ? ? ? monday
?--------- ? ? ? ? ? monday.1
[tdh@adept snapshot]> cd monday
monday: Object is remote.
[tdh@adept snapshot]> uname
Linux
[tdh@adept snapshot]> mount | grep wont
wont:/export/zfs on /net/wont/export/zfs type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
wont:/export/zfs/tdh on /net/wont/export/zfs/tdh type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
wont:/export/zfs/nfsv4 on /net/wont/export/zfs/nfsv4 type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
wont:/export/zfs/nfsv3 on /net/wont/export/zfs/nfsv3 type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
wont:/export/zfs/nfsv2 on /net/wont/export/zfs/nfsv2 type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
[tdh@adept snapshot]> mount | grep wont
wont:/export/zfs on /net/wont/export/zfs type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
wont:/export/zfs/tdh on /net/wont/export/zfs/tdh type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
wont:/export/zfs/nfsv4 on /net/wont/export/zfs/nfsv4 type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
wont:/export/zfs/nfsv3 on /net/wont/export/zfs/nfsv3 type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
wont:/export/zfs/nfsv2 on /net/wont/export/zfs/nfsv2 type nfs
(rw,nosuid,nodev,hard,intr,addr=192.168.2.105)
NFSv4 works:
[tdh@adept snapshot]> sudo mkdir -p /nfsv4/wont/export/zfs/nfsv4
[tdh@adept snapshot]> sudo mount -t nfs4 wont:/export/zfs/nfsv4
/nfsv4/wont/export/zfs/nfsv4
[tdh@adept snapshot]> cd !$
cd /nfsv4/wont/export/zfs/nfsv4
[tdh@adept nfsv4]> ls -la
total 16
drwxr-xr-x 2 nfsv4 nobody 4 Mar 20 01:44 .
drwxr-xr-x 3 root root 4096 Mar 21 18:01 ..
-rw-r--r-- 1 nfsv4 nobody 0 Mar 20 01:04 it
dr-xr-xr-x 3 root root 3 Mar 21 17:55 .zfs
-rw-r--r-- 1 nfsv4 nobody 13867 Mar 20 01:47 zfs.txt
[tdh@adept nfsv4]> cd .zfs
[tdh@adept .zfs]> ls -la
total 1
dr-xr-xr-x 3 root root 3 Mar 21 17:56 .
drwxr-xr-x 2 nfsv4 nobody 4 Mar 20 01:44 ..
dr-xr-xr-x 4 root root 4 Mar 21 17:56 snapshot
[tdh@adept .zfs]> cd snapshot/
[tdh@adept snapshot]> ls -la
total 2
dr-xr-xr-x 4 root root 4 Mar 21 17:56 .
dr-xr-xr-x 3 root root 3 Mar 21 17:56 ..
drwxr-xr-x 2 nfsv4 nobody 4 Mar 20 01:38 monday
drwxr-xr-x 2 nfsv4 nobody 4 Mar 20 01:44 monday.1
[tdh@adept snapshot]> cd monday
[tdh@adept monday]> ls -la
total 11
drwxr-xr-x 2 nfsv4 nobody 4 Mar 20 01:38 .
dr-xr-xr-x 4 root root 4 Mar 21 17:56 ..
-rw-r--r-- 1 nfsv4 nobody 0 Mar 20 01:04 it
-rw-r--r-- 1 nfsv4 nobody 11808 Mar 20 01:38 zfs.txt
--
Tom Haynes
tdh at sun dot com
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
79
From:
US
Registered:
6/7/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 4:07 PM
in response to: ormandj
|
|
David;
This smells like a readdir loop, i thought that issues with snapshots
(and v4) were resolved and included in build 33 (which represents
SX 03/06) ..
Would you be so kind to collect some information for us ?
The version of your FreeBSD Client:
Capture all the packets (tcpdump/snoop/ethereal) between the client
and server:
on the FreeBSD client that would be
"tcpdump -i <if_name> -w tcp.raw host server2"
at the server run this little "D"-ness ..
#!/usr/sbin/dtrace -Fs
rfs4_op_readdir:entry
{
self->start = vtimestamp;
}
rfs4_op_readdir:return
{
self->start = 0;
}
*_readdir:entry
/self->start/
{
self->uio = (struct uio *)arg1;
printf("vp - %p; ", arg0);
printf(" offset = %x\n", self->uio->_uio_offset._f);
trace(vtimestamp - self->start);
}
-- Robert..
On Mar 21, 2006, at 4:42 PM, David J. Orman wrote:
> Hi,
>
> After having gotten the other issue sorted, now I'm running into a
> new one!
>
> I made a snapshot of a user's home directory. I can browse to it
> just fine
> on the Solaris machine running zfs.
>
> # pwd
> /export/zfs/chuangj/.zfs/snapshot/03212006
> #
>
> # zfs list
> NAME USED AVAIL REFER MOUNTPOINT
> users 18.5G 166G 98.5K /users
> users/home 18.5G 166G 99.5K /export/zfs
> users/home/chuangj 3.17G 166G 3.17G /export/zfs/chuangj
> users/home/chuangj@03212006 498K - 3.17G -
> #
>
> However, on the FreeBSD machine, it's not working.
>
> Filesystem Size Used Avail Capacity
> Mounted on
> server2:export/zfs/chuangj 169G 3.2G 166G 2%
> /usr/home/chuangj
>
> %pwd
> /usr/home/chuangj/.zfs/snapshot
> %ls
> 03212006
> %cd 03212006
> 03212006: Too many levels of remote in path.
> %
> %ls -als
> ls: 03212006: Too many levels of remote in path
> total 0
> 0 dr-xr-xr-x 3 root wheel 3 Mar 21 12:41 .
> 0 dr-xr-xr-x 3 root wheel 3 Mar 21 12:41 ..
> %
>
> Reading up some, I've seen some references to not being able to
> mount the
> filesystem because it's already mounted. However, this didn't make
> clear
> what I'm doing wrong. :) Once again, I ask for your kind assistance!
>
> Cheers,
> David
>
> PS - I'm really loving ZFS so far, this is FUN!
>
> _______________________________________________
> nfs-discuss mailing list
> nfs-discuss at opensolaris dot org
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
192
From:
Registered:
2/21/06
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 4:24 PM
in response to: rbg
|
|
Ok, I wasn't too clear on the dtrace bit, as I've never used it, but I
stuck that script you posted into a file called d.sh and ran it. Then I
started tcpdump as you specified on the client machine. I then attempted
to cd to that .zfs/snapshot/03212006 directory. I hope this is what you
wanted. :)
>From the dtrace script:
# ./d.sh
dtrace: script './d.sh' matched 25 probes
^C
Got no output. I let it sit for a bit but nothing came out, so I ^C'd it.
FreeBSD client:
http://corenode.com/~ormandj/files/tcp.raw.gz
Is this what you needed? Let me know if I can help any other way. BTW - is
FreeBSD connecting as NFSv4? It sounded like you're going to figure out
the client version from the output of tcpdump, and I'd be curious to know
myself. Thanks!
David
> David;
>
> This smells like a readdir loop, i thought that issues with snapshots
> (and v4) were resolved and included in build 33 (which represents
> SX 03/06) ..
>
> Would you be so kind to collect some information for us ?
>
> The version of your FreeBSD Client:
>
> Capture all the packets (tcpdump/snoop/ethereal) between the client
> and server:
> on the FreeBSD client that would be
> "tcpdump -i <if_name> -w tcp.raw host server2"
>
> at the server run this little "D"-ness ..
>
> #!/usr/sbin/dtrace -Fs
>
> rfs4_op_readdir:entry
> {
> self->start = vtimestamp;
> }
>
> rfs4_op_readdir:return
> {
> self->start = 0;
> }
>
> *_readdir:entry
> /self->start/
> {
> self->uio = (struct uio *)arg1;
>
> printf("vp - %p; ", arg0);
> printf(" offset = %x\n", self->uio->_uio_offset._f);
>
> trace(vtimestamp - self->start);
> }
>
>
> -- Robert..
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
79
From:
US
Registered:
6/7/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 4:33 PM
in response to: ormandj
|
|
David..
I think that Tom was right in that you are seeing the little issue
with v4 vs v4... Try your mount with NFSv4 as Tom demonstrated :-
"mount -t nfs4 wont:/export/zfs/nfsv4"
Seeing nothing out of Dtrace reenforces the threroy of v3.. since
I didn't 'hook' the rfs3 readdir.. (i'll peek at the dump)
-- Robert..
On Mar 21, 2006, at 6:24 PM, David J. Orman wrote:
> Ok, I wasn't too clear on the dtrace bit, as I've never used it, but I
> stuck that script you posted into a file called d.sh and ran it.
> Then I
> started tcpdump as you specified on the client machine. I then
> attempted
> to cd to that .zfs/snapshot/03212006 directory. I hope this is what
> you
> wanted. :)
>
> From the dtrace script:
>
> # ./d.sh
> dtrace: script './d.sh' matched 25 probes
> ^C
>
> Got no output. I let it sit for a bit but nothing came out, so I
> ^C'd it.
>
> FreeBSD client:
>
> http://corenode.com/~ormandj/files/tcp.raw.gz
>
> Is this what you needed? Let me know if I can help any other way.
> BTW - is
> FreeBSD connecting as NFSv4? It sounded like you're going to figure
> out
> the client version from the output of tcpdump, and I'd be curious
> to know
> myself. Thanks!
>
> David
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
192
From:
Registered:
2/21/06
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 4:40 PM
in response to: rbg
|
|
Ok, I found mount_nfs4 hidden in the freebsd man pages. I mounted via
NFS4, and I could cd into the snapshot dirs. However, one slight problem.
All the files were owned/grouped funky:
Mounted via NFSv4:
%ls -als
total 1315
0 drwxrwxrwx 15 4294967294 4294967294 37 Feb 28 14:45 .
0 drwxr-xr-x 13 4294967294 4294967294 25 Feb 28 14:45 ..
4 -rwxr-xr-x 1 4294967294 4294967294 3584 Feb 28 14:52 .htaccess
55 -rwxr-xr-x 1 4294967294 4294967294 56216 Feb 25 23:26 CHANGELOG.php
3 -rw-r--r-- 1 4294967294 4294967294 3429 Jan 4 03:14 COPYRIGHT.php
0 -rw-r--r-- 1 4294967294 4294967294 0 Dec 1 13:11
GOOGLE9c422d3cbcc625d8.html
.... etc
When mounted with NFSv3:
server1# ls -als
total 1651
3 drwxrwxrwx 15 corenode corenode 37 Feb 28 14:45 .
2 drwxr-xr-x 13 corenode corenode 25 Feb 28 14:45 ..
3 -rwxr-xr-x 1 corenode corenode 3584 Feb 28 14:52 .htaccess
39 -rwxr-xr-x 1 corenode corenode 56216 Feb 25 23:26 CHANGELOG.php
3 -rw-r--r-- 1 corenode corenode 3429 Jan 4 03:14 COPYRIGHT.php
1 -rw-r--r-- 1 corenode corenode 0 Dec 1 13:11
GOOGLE9c422d3cbcc625d8.html
etc....
Needless to say, it looks like I won't be using v4, at least not with
freebsd. I used rsync -aP to migrate the files onto zfs, while it was
mounted NFSv3. I then unmounted the share, and remounted it using NFSv4.
Cheers,
David
> David..
>
> I think that Tom was right in that you are seeing the little issue
> with v4 vs v4... Try your mount with NFSv4 as Tom demonstrated :-
> "mount -t nfs4 wont:/export/zfs/nfsv4"
>
> Seeing nothing out of Dtrace reenforces the threroy of v3.. since
> I didn't 'hook' the rfs3 readdir.. (i'll peek at the dump)
>
> -- Robert..
>
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
370
From:
US
Registered:
6/13/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4
w/ FreeBSD client
Posted:
Mar 21, 2006 5:46 PM
in response to: ormandj
|
|
David J. Orman wrote:
>Ok, I found mount_nfs4 hidden in the freebsd man pages. I mounted via
>NFS4, and I could cd into the snapshot dirs. However, one slight problem.
>All the files were owned/grouped funky:
>
>
>
My guess is that you need to change match your ID domain on the server
and client. In solaris, we call the ID domain NFSMAPID_DOMAIN - not
sure what BSD calls it, check out:
http://blogs.sun.com/nfsv4
http://blogs.sun.com/roller/page/erickustarz/20050128
eric
>Mounted via NFSv4:
>
>%ls -als
>total 1315
> 0 drwxrwxrwx 15 4294967294 4294967294 37 Feb 28 14:45 .
> 0 drwxr-xr-x 13 4294967294 4294967294 25 Feb 28 14:45 ..
> 4 -rwxr-xr-x 1 4294967294 4294967294 3584 Feb 28 14:52 .htaccess
> 55 -rwxr-xr-x 1 4294967294 4294967294 56216 Feb 25 23:26 CHANGELOG.php
> 3 -rw-r--r-- 1 4294967294 4294967294 3429 Jan 4 03:14 COPYRIGHT.php
> 0 -rw-r--r-- 1 4294967294 4294967294 0 Dec 1 13:11
>GOOGLE9c422d3cbcc625d8.html
>.... etc
>
>When mounted with NFSv3:
>
>server1# ls -als
>total 1651
> 3 drwxrwxrwx 15 corenode corenode 37 Feb 28 14:45 .
> 2 drwxr-xr-x 13 corenode corenode 25 Feb 28 14:45 ..
> 3 -rwxr-xr-x 1 corenode corenode 3584 Feb 28 14:52 .htaccess
> 39 -rwxr-xr-x 1 corenode corenode 56216 Feb 25 23:26 CHANGELOG.php
> 3 -rw-r--r-- 1 corenode corenode 3429 Jan 4 03:14 COPYRIGHT.php
> 1 -rw-r--r-- 1 corenode corenode 0 Dec 1 13:11
>GOOGLE9c422d3cbcc625d8.html
>etc....
>
>Needless to say, it looks like I won't be using v4, at least not with
>freebsd. I used rsync -aP to migrate the files onto zfs, while it was
>mounted NFSv3. I then unmounted the share, and remounted it using NFSv4.
>
>Cheers,
>David
>
>
>
>>David..
>>
>>I think that Tom was right in that you are seeing the little issue
>>with v4 vs v4... Try your mount with NFSv4 as Tom demonstrated :-
>> "mount -t nfs4 wont:/export/zfs/nfsv4"
>>
>>Seeing nothing out of Dtrace reenforces the threroy of v3.. since
>>I didn't 'hook' the rfs3 readdir.. (i'll peek at the dump)
>>
>>-- Robert..
>>
>>
>>
>
>
>_______________________________________________
>nfs-discuss mailing list
>nfs-discuss at opensolaris dot org
>
>
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
79
From:
US
Registered:
6/7/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 8:25 PM
in response to: ormandj
|
|
David;
Looks like the mapping of user/group strings to the numeric values is
performed
by a combination of the kernel and a user daemon called nfsuserd.
Looking at
the FreeBSD 6.0 code from Ricks FTP site the kernel is initially
seeded with
user/groups and then performs 'upcalls' to resolve cache misses.
nfsuserd will
use the doamin portion that is returned from the gethostname() call,
or whatever
you specify via the command line using the '-domain' option. If you
used the
start.nfs script from nfsv4utils the domain should have been printed
to stderr
when nfsuserd started something like "nfsuserd:
domain=corenode.com ...."
On the server 'cat /var/run/nfs4_domain' this will tell you what the
Solaris
server thinks the correct domain is... if they match, then "beeep,
houston
we have a problem" .. If they do not match, then you have the choice of
resetting the server or the client, or for now make 'em both the same
by :-
adding the -domain option to nfsuserd in the start.nfs script
----------------
#!/bin/sh
./nfsuserd/nfsuserd -verbose -domain corenode.com
...
-----------------
and on the server:
edit /etc/default/nfs replacing
"#NFSMAPID_DOMAIN=domain"
with
"NFSMAPID_DOMAIN=corenode.com"
Then execute:
# svcadm restart svc:/network/nfs/mapid
I'm not sure how to restart the nfsuserd for freebsd, i'd guess that
you could kill the current nfsuserd then restart it by hand with this
command-line
# nfsuserd -verbose -force -domain corenode.com
again, for the FreeBSD stuff, i've just looked at the code .. Let us
know
how you get on, and if you still see problem we'll need another tcpdump
of the traffic...
-- Robert..
On Mar 21, 2006, at 6:40 PM, David J. Orman wrote:
> Ok, I found mount_nfs4 hidden in the freebsd man pages. I mounted via
> NFS4, and I could cd into the snapshot dirs. However, one slight
> problem.
> All the files were owned/grouped funky:
>
> Mounted via NFSv4:
>
> %ls -als
> total 1315
> 0 drwxrwxrwx 15 4294967294 4294967294 37 Feb 28 14:45 .
> 0 drwxr-xr-x 13 4294967294 4294967294 25 Feb 28 14:45 ..
> 4 -rwxr-xr-x 1 4294967294 4294967294 3584 Feb 28
> 14:52 .htaccess
> 55 -rwxr-xr-x 1 4294967294 4294967294 56216 Feb 25 23:26
> CHANGELOG.php
> 3 -rw-r--r-- 1 4294967294 4294967294 3429 Jan 4 03:14
> COPYRIGHT.php
> 0 -rw-r--r-- 1 4294967294 4294967294 0 Dec 1 13:11
> GOOGLE9c422d3cbcc625d8.html
> .... etc
>
> When mounted with NFSv3:
>
> server1# ls -als
> total 1651
> 3 drwxrwxrwx 15 corenode corenode 37 Feb 28 14:45 .
> 2 drwxr-xr-x 13 corenode corenode 25 Feb 28 14:45 ..
> 3 -rwxr-xr-x 1 corenode corenode 3584 Feb 28 14:52 .htaccess
> 39 -rwxr-xr-x 1 corenode corenode 56216 Feb 25 23:26
> CHANGELOG.php
> 3 -rw-r--r-- 1 corenode corenode 3429 Jan 4 03:14
> COPYRIGHT.php
> 1 -rw-r--r-- 1 corenode corenode 0 Dec 1 13:11
> GOOGLE9c422d3cbcc625d8.html
> etc....
>
> Needless to say, it looks like I won't be using v4, at least not with
> freebsd. I used rsync -aP to migrate the files onto zfs, while it was
> mounted NFSv3. I then unmounted the share, and remounted it using
> NFSv4.
>
> Cheers,
> David
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
192
From:
Registered:
2/21/06
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 8:44 PM
in response to: rbg
|
|
I checked into that. There is no nfsv4utils to be found, nor nfsuserd
in fbsd6.
This is the same thing, from a mailing list:
http://groups.google.com/group/lucky.freebsd.current/browse_frm/
thread/8704ebdb69cd23d0/5a6804849fc85a9b?lnk=st&q=freebsd+nfsv4+domain
+id&rnum=3&hl=en#5a6804849fc85a9b
The first few posts are pointless (trying to use old tools...) but
after that he/she ends up arriving with the same fate as me. Google
didn't turn up any ways to specify the domain id in freebsd either.
I'm using "mount_nfs4" to mount as nfs4.
http://www.citi.umich.edu/projects/nfsv4/bsd-nfsv4/FreeBSD-mount_nfs4/
Let me know if I can help out with more information, I really
appreciate all of your help for me. :)
Cheers,
David
On Mar 21, 2006, at 6:25 PM, Robert Gordon wrote:
>
> David;
>
> Looks like the mapping of user/group strings to the numeric values
> is performed
> by a combination of the kernel and a user daemon called nfsuserd.
> Looking at
> the FreeBSD 6.0 code from Ricks FTP site the kernel is initially
> seeded with
> user/groups and then performs 'upcalls' to resolve cache misses.
> nfsuserd will
> use the doamin portion that is returned from the gethostname()
> call, or whatever
> you specify via the command line using the '-domain' option. If you
> used the
> start.nfs script from nfsv4utils the domain should have been
> printed to stderr
> when nfsuserd started something like "nfsuserd:
> domain=corenode.com ...."
>
> On the server 'cat /var/run/nfs4_domain' this will tell you what
> the Solaris
> server thinks the correct domain is... if they match, then "beeep,
> houston
> we have a problem" .. If they do not match, then you have the
> choice of
> resetting the server or the client, or for now make 'em both the
> same by :-
>
> adding the -domain option to nfsuserd in the start.nfs script
> ----------------
> #!/bin/sh
> ./nfsuserd/nfsuserd -verbose -domain corenode.com
> ...
> -----------------
>
> and on the server:
>
> edit /etc/default/nfs replacing
> "#NFSMAPID_DOMAIN=domain"
> with
> "NFSMAPID_DOMAIN=corenode.com"
>
> Then execute:
>
> # svcadm restart svc:/network/nfs/mapid
>
> I'm not sure how to restart the nfsuserd for freebsd, i'd guess that
> you could kill the current nfsuserd then restart it by hand with this
> command-line
>
> # nfsuserd -verbose -force -domain corenode.com
>
> again, for the FreeBSD stuff, i've just looked at the code .. Let
> us know
> how you get on, and if you still see problem we'll need another
> tcpdump
> of the traffic...
>
> -- Robert..
>
> On Mar 21, 2006, at 6:40 PM, David J. Orman wrote:
>> Ok, I found mount_nfs4 hidden in the freebsd man pages. I mounted via
>> NFS4, and I could cd into the snapshot dirs. However, one slight
>> problem.
>> All the files were owned/grouped funky:
>>
>> Mounted via NFSv4:
>>
>> %ls -als
>> total 1315
>> 0 drwxrwxrwx 15 4294967294 4294967294 37 Feb 28 14:45 .
>> 0 drwxr-xr-x 13 4294967294 4294967294 25 Feb 28 14:45 ..
>> 4 -rwxr-xr-x 1 4294967294 4294967294 3584 Feb 28
>> 14:52 .htaccess
>> 55 -rwxr-xr-x 1 4294967294 4294967294 56216 Feb 25 23:26
>> CHANGELOG.php
>> 3 -rw-r--r-- 1 4294967294 4294967294 3429 Jan 4 03:14
>> COPYRIGHT.php
>> 0 -rw-r--r-- 1 4294967294 4294967294 0 Dec 1 13:11
>> GOOGLE9c422d3cbcc625d8.html
>> .... etc
>>
>> When mounted with NFSv3:
>>
>> server1# ls -als
>> total 1651
>> 3 drwxrwxrwx 15 corenode corenode 37 Feb 28 14:45 .
>> 2 drwxr-xr-x 13 corenode corenode 25 Feb 28 14:45 ..
>> 3 -rwxr-xr-x 1 corenode corenode 3584 Feb 28 14:52 .htaccess
>> 39 -rwxr-xr-x 1 corenode corenode 56216 Feb 25 23:26
>> CHANGELOG.php
>> 3 -rw-r--r-- 1 corenode corenode 3429 Jan 4 03:14
>> COPYRIGHT.php
>> 1 -rw-r--r-- 1 corenode corenode 0 Dec 1 13:11
>> GOOGLE9c422d3cbcc625d8.html
>> etc....
>>
>> Needless to say, it looks like I won't be using v4, at least not with
>> freebsd. I used rsync -aP to migrate the files onto zfs, while it was
>> mounted NFSv3. I then unmounted the share, and remounted it using
>> NFSv4.
>>
>> Cheers,
>> David
>
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
79
From:
US
Registered:
6/7/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 10:03 PM
in response to: ormandj
|
|
!! BANG !! (that was my head exploding)
Digging a little deeper, it turns out that where i say FreeBSD,
you should replace with OpenBSD..
On Mar 21, 2006, at 10:44 PM, David J. Orman wrote:
> I checked into that. There is no nfsv4utils to be found, nor
> nfsuserd in fbsd6.
>
> This is the same thing, from a mailing list:
>
> http://groups.google.com/group/lucky.freebsd.current/browse_frm/
> thread/8704ebdb69cd23d0/5a6804849fc85a9b?lnk=st&q=freebsd+nfsv4
> +domain+id&rnum=3&hl=en#5a6804849fc85a9b
>
> The first few posts are pointless (trying to use old tools...) but
> after that he/she ends up arriving with the same fate as me. Google
> didn't turn up any ways to specify the domain id in freebsd either.
>
> I'm using "mount_nfs4" to mount as nfs4.
> http://www.citi.umich.edu/projects/nfsv4/bsd-nfsv4/FreeBSD-mount_nfs4/
>
> Let me know if I can help out with more information, I really
> appreciate all of your help for me. :)
>
> Cheers,
> David
>
> On Mar 21, 2006, at 6:25 PM, Robert Gordon wrote:
>
>>
>> David;
>>
>> Looks like the mapping of user/group strings to the numeric values
>> is performed
>> by a combination of the kernel and a user daemon called nfsuserd.
>> Looking at
>> the FreeBSD 6.0 code from Ricks FTP site the kernel is initially
>> seeded with
>> user/groups and then performs 'upcalls' to resolve cache misses.
>> nfsuserd will
>> use the doamin portion that is returned from the gethostname()
>> call, or whatever
>> you specify via the command line using the '-domain' option. If
>> you used the
>> start.nfs script from nfsv4utils the domain should have been
>> printed to stderr
>> when nfsuserd started something like "nfsuserd:
>> domain=corenode.com ...."
>>
>> On the server 'cat /var/run/nfs4_domain' this will tell you what
>> the Solaris
>> server thinks the correct domain is... if they match, then "beeep,
>> houston
>> we have a problem" .. If they do not match, then you have the
>> choice of
>> resetting the server or the client, or for now make 'em both the
>> same by :-
>>
>> adding the -domain option to nfsuserd in the start.nfs script
>> ----------------
>> #!/bin/sh
>> ./nfsuserd/nfsuserd -verbose -domain corenode.com
>> ...
>> -----------------
>>
>> and on the server:
>>
>> edit /etc/default/nfs replacing
>> "#NFSMAPID_DOMAIN=domain"
>> with
>> "NFSMAPID_DOMAIN=corenode.com"
>>
>> Then execute:
>>
>> # svcadm restart svc:/network/nfs/mapid
>>
>> I'm not sure how to restart the nfsuserd for freebsd, i'd guess that
>> you could kill the current nfsuserd then restart it by hand with this
>> command-line
>>
>> # nfsuserd -verbose -force -domain corenode.com
>>
>> again, for the FreeBSD stuff, i've just looked at the code .. Let
>> us know
>> how you get on, and if you still see problem we'll need another
>> tcpdump
>> of the traffic...
>>
>> -- Robert..
>>
>> On Mar 21, 2006, at 6:40 PM, David J. Orman wrote:
>>> Ok, I found mount_nfs4 hidden in the freebsd man pages. I mounted
>>> via
>>> NFS4, and I could cd into the snapshot dirs. However, one slight
>>> problem.
>>> All the files were owned/grouped funky:
>>>
>>> Mounted via NFSv4:
>>>
>>> %ls -als
>>> total 1315
>>> 0 drwxrwxrwx 15 4294967294 4294967294 37 Feb 28 14:45 .
>>> 0 drwxr-xr-x 13 4294967294 4294967294 25 Feb 28 14:45 ..
>>> 4 -rwxr-xr-x 1 4294967294 4294967294 3584 Feb 28
>>> 14:52 .htaccess
>>> 55 -rwxr-xr-x 1 4294967294 4294967294 56216 Feb 25 23:26
>>> CHANGELOG.php
>>> 3 -rw-r--r-- 1 4294967294 4294967294 3429 Jan 4 03:14
>>> COPYRIGHT.php
>>> 0 -rw-r--r-- 1 4294967294 4294967294 0 Dec 1 13:11
>>> GOOGLE9c422d3cbcc625d8.html
>>> .... etc
>>>
>>> When mounted with NFSv3:
>>>
>>> server1# ls -als
>>> total 1651
>>> 3 drwxrwxrwx 15 corenode corenode 37 Feb 28 14:45 .
>>> 2 drwxr-xr-x 13 corenode corenode 25 Feb 28 14:45 ..
>>> 3 -rwxr-xr-x 1 corenode corenode 3584 Feb 28
>>> 14:52 .htaccess
>>> 39 -rwxr-xr-x 1 corenode corenode 56216 Feb 25 23:26
>>> CHANGELOG.php
>>> 3 -rw-r--r-- 1 corenode corenode 3429 Jan 4 03:14
>>> COPYRIGHT.php
>>> 1 -rw-r--r-- 1 corenode corenode 0 Dec 1 13:11
>>> GOOGLE9c422d3cbcc625d8.html
>>> etc....
>>>
>>> Needless to say, it looks like I won't be using v4, at least not
>>> with
>>> freebsd. I used rsync -aP to migrate the files onto zfs, while it
>>> was
>>> mounted NFSv3. I then unmounted the share, and remounted it using
>>> NFSv4.
>>>
>>> Cheers,
>>> David
>>
>
> _______________________________________________
> nfs-discuss mailing list
> nfs-discuss at opensolaris dot org
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
192
From:
Registered:
2/21/06
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 10:11 PM
in response to: rbg
|
|
*BANG* That's the sound of my credit card exploding, after hearing
about it's soon-to-be-demise buying HCL stuff to continue evaluating
Solaris, and it's not-too-far-off complete and utter obliteration at
the Sun store. Thanks for the help, it seems nfsv4 is Sun-territory
(and maybe a little linux thrown into the mix! Just won't be in MY
mix. :P)
Cheers,
David
On Mar 21, 2006, at 8:03 PM, Robert Gordon wrote:
>
>
> !! BANG !! (that was my head exploding)
>
> Digging a little deeper, it turns out that where i say FreeBSD,
> you should replace with OpenBSD..
>
> On Mar 21, 2006, at 10:44 PM, David J. Orman wrote:
>
>> I checked into that. There is no nfsv4utils to be found, nor
>> nfsuserd in fbsd6.
>>
>> This is the same thing, from a mailing list:
>>
>> http://groups.google.com/group/lucky.freebsd.current/browse_frm/
>> thread/8704ebdb69cd23d0/5a6804849fc85a9b?lnk=st&q=freebsd+nfsv4
>> +domain+id&rnum=3&hl=en#5a6804849fc85a9b
>>
>> The first few posts are pointless (trying to use old tools...) but
>> after that he/she ends up arriving with the same fate as me.
>> Google didn't turn up any ways to specify the domain id in freebsd
>> either.
>>
>> I'm using "mount_nfs4" to mount as nfs4.
>> http://www.citi.umich.edu/projects/nfsv4/bsd-nfsv4/FreeBSD-
>> mount_nfs4/
>>
>> Let me know if I can help out with more information, I really
>> appreciate all of your help for me. :)
>>
>> Cheers,
>> David
>>
>> On Mar 21, 2006, at 6:25 PM, Robert Gordon wrote:
>>
>>>
>>> David;
>>>
>>> Looks like the mapping of user/group strings to the numeric
>>> values is performed
>>> by a combination of the kernel and a user daemon called nfsuserd.
>>> Looking at
>>> the FreeBSD 6.0 code from Ricks FTP site the kernel is initially
>>> seeded with
>>> user/groups and then performs 'upcalls' to resolve cache misses.
>>> nfsuserd will
>>> use the doamin portion that is returned from the gethostname()
>>> call, or whatever
>>> you specify via the command line using the '-domain' option. If
>>> you used the
>>> start.nfs script from nfsv4utils the domain should have been
>>> printed to stderr
>>> when nfsuserd started something like "nfsuserd:
>>> domain=corenode.com ...."
>>>
>>> On the server 'cat /var/run/nfs4_domain' this will tell you what
>>> the Solaris
>>> server thinks the correct domain is... if they match, then
>>> "beeep, houston
>>> we have a problem" .. If they do not match, then you have the
>>> choice of
>>> resetting the server or the client, or for now make 'em both the
>>> same by :-
>>>
>>> adding the -domain option to nfsuserd in the start.nfs script
>>> ----------------
>>> #!/bin/sh
>>> ./nfsuserd/nfsuserd -verbose -domain corenode.com
>>> ...
>>> -----------------
>>>
>>> and on the server:
>>>
>>> edit /etc/default/nfs replacing
>>> "#NFSMAPID_DOMAIN=domain"
>>> with
>>> "NFSMAPID_DOMAIN=corenode.com"
>>>
>>> Then execute:
>>>
>>> # svcadm restart svc:/network/nfs/mapid
>>>
>>> I'm not sure how to restart the nfsuserd for freebsd, i'd guess that
>>> you could kill the current nfsuserd then restart it by hand with
>>> this
>>> command-line
>>>
>>> # nfsuserd -verbose -force -domain corenode.com
>>>
>>> again, for the FreeBSD stuff, i've just looked at the code .. Let
>>> us know
>>> how you get on, and if you still see problem we'll need another
>>> tcpdump
>>> of the traffic...
>>>
>>> -- Robert..
>>>
>>> On Mar 21, 2006, at 6:40 PM, David J. Orman wrote:
>>>> Ok, I found mount_nfs4 hidden in the freebsd man pages. I
>>>> mounted via
>>>> NFS4, and I could cd into the snapshot dirs. However, one slight
>>>> problem.
>>>> All the files were owned/grouped funky:
>>>>
>>>> Mounted via NFSv4:
>>>>
>>>> %ls -als
>>>> total 1315
>>>> 0 drwxrwxrwx 15 4294967294 4294967294 37 Feb 28 14:45 .
>>>> 0 drwxr-xr-x 13 4294967294 4294967294 25 Feb 28 14:45 ..
>>>> 4 -rwxr-xr-x 1 4294967294 4294967294 3584 Feb 28
>>>> 14:52 .htaccess
>>>> 55 -rwxr-xr-x 1 4294967294 4294967294 56216 Feb 25 23:26
>>>> CHANGELOG.php
>>>> 3 -rw-r--r-- 1 4294967294 4294967294 3429 Jan 4 03:14
>>>> COPYRIGHT.php
>>>> 0 -rw-r--r-- 1 4294967294 4294967294 0 Dec 1 13:11
>>>> GOOGLE9c422d3cbcc625d8.html
>>>> .... etc
>>>>
>>>> When mounted with NFSv3:
>>>>
>>>> server1# ls -als
>>>> total 1651
>>>> 3 drwxrwxrwx 15 corenode corenode 37 Feb 28 14:45 .
>>>> 2 drwxr-xr-x 13 corenode corenode 25 Feb 28 14:45 ..
>>>> 3 -rwxr-xr-x 1 corenode corenode 3584 Feb 28
>>>> 14:52 .htaccess
>>>> 39 -rwxr-xr-x 1 corenode corenode 56216 Feb 25 23:26
>>>> CHANGELOG.php
>>>> 3 -rw-r--r-- 1 corenode corenode 3429 Jan 4 03:14
>>>> COPYRIGHT.php
>>>> 1 -rw-r--r-- 1 corenode corenode 0 Dec 1 13:11
>>>> GOOGLE9c422d3cbcc625d8.html
>>>> etc....
>>>>
>>>> Needless to say, it looks like I won't be using v4, at least not
>>>> with
>>>> freebsd. I used rsync -aP to migrate the files onto zfs, while
>>>> it was
>>>> mounted NFSv3. I then unmounted the share, and remounted it
>>>> using NFSv4.
>>>>
>>>> Cheers,
>>>> David
>>>
>>
>> _______________________________________________
>> nfs-discuss mailing list
>> nfs-discuss at opensolaris dot org
>
> _______________________________________________
> nfs-discuss mailing list
> nfs-discuss at opensolaris dot org
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
466
From:
US
Registered:
12/21/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/
FreeBSD client
Posted:
Mar 21, 2006 10:45 PM
in response to: rbg
|
|
Robert Gordon wrote On 03/22/06 00:03,:
>
> !! BANG !! (that was my head exploding)
>
> Digging a little deeper, it turns out that where i say FreeBSD,
> you should replace with OpenBSD..
>
> On Mar 21, 2006, at 10:44 PM, David J. Orman wrote:
>
>
>>I checked into that. There is no nfsv4utils to be found, nor
>>nfsuserd in fbsd6.
>>
>>This is the same thing, from a mailing list:
>>
>>http://groups.google.com/group/lucky.freebsd.current/browse_frm/
>>thread/8704ebdb69cd23d0/5a6804849fc85a9b?lnk=st&q=freebsd+nfsv4
>>+domain+id&rnum=3&hl=en#5a6804849fc85a9b
>>
>>The first few posts are pointless (trying to use old tools...) but
>>after that he/she ends up arriving with the same fate as me. Google
>>didn't turn up any ways to specify the domain id in freebsd either.
>>
>>I'm using "mount_nfs4" to mount as nfs4.
>>http://www.citi.umich.edu/projects/nfsv4/bsd-nfsv4/FreeBSD-mount_nfs4/
>>
>>Let me know if I can help out with more information, I really
>>appreciate all of your help for me. :)
>>
>>Cheers,
>>David
>>
You could look at the idmapd manpage, but it doesn't exist. The source
code is there, but I think there is a gap here.
I couldn't find the daemon anywhere on disk after I installed FreeBSD
6.0. Also, once I remembered it was 'ps -aux', I didn't see it running.
Anyway, the original bug which causes the NFSv3 <-> ZFS interaction has
been fixed and should be out in OpenSolaris sooner than later.
--
Tom Haynes
tdh at sun dot com
_______________________________________________
nfs-discuss mailing list
nfs-discuss at opensolaris dot org
|
|
|
|
|
Posts:
383
From:
GB
Registered:
10/27/05
|
|
|
|
Re: Problem accessing snapshot dirs over NFSv4 w/FreeBSD client
Posted:
Mar 22, 2006 5:58 AM
in response to: tdh
|
|
> You could look at the idmapd manpage, but it doesn't
> exist. The source code is there, but I think there is a gap here.
>
> I couldn't find the daemon anywhere on disk after I
> installed FreeBSD 6.0. Also, once I remembered it was 'ps -aux', I
> didn't see it running.
We don't install it. Our NFSv4 client is currently considered "experimental", though quite why we only give you half of it, I couldn't rightly say at the minute. I'll get on to some people.
However, if you download the source code, it should build and work.
|
|
|
|
|
