NFS write failed for server : error 27 (RPC: Received disconnect from remote) - Solaris 9

[vpek...@gmail.com]

I have an issue with an nfs mount that is mounted and i can list the items under the dir. However when i try to write to it, it hangs and shows this error: NFS server *** not responding still trying. This only shows up when i try to write to it and looks like its only happening on solaris 9. I have mounted this nfs share elsewhere on the same network without issues on sol10 and redhat linux. This is intermittent but recently every time I check , I have been getting this error and cannot write to it. So i don't believe its a NAS OR network issue but more related to solaris 9 deficiency of some sort? Can someone please advise

[YTC#1]

On 26/06/2018 20:12, vpek...@gmail.com wrote:
> I have an issue with an nfs mount that is mounted and i can list the items under the dir. However when i try to write to it, it hangs and shows this error: NFS server *** not responding still trying. This only shows up when i try to write to it and looks like its only happening on solaris 9. I have mounted this nfs share elsewhere on the same network without issues on sol10 and redhat linux. This is intermittent but recently every time I check , I have been getting this error and cannot write to it. So i don't believe its a NAS OR network issue but more related to solaris 9 deficiency of some sort? Can someone please advise
>

I think we need some more info here.

On the S9 client, how is the FS mounted ? (/etc/dfs/dfstab ??? if so,
contents)

Does the NFS sever ping from the client ?

Are there any other errors on the client (/var/adm/messages) that might
give a clue.

Are all the NFS services running on the client ? (So, old school,
/etc/rc3.d/S* )

Are the client and the server using the same NFS version ?


--
Bruce Porter
"The internet is a huge and diverse community but mainly friendly"
http://ytc1.blogspot.co.uk/
There *is* an alternative! http://www.openoffice.org/

[vpek...@gmail.com]

from vfstab:
tfnas9:/vol/dbsarchive - /dbsarchive nfs - yes bg,hard,retry=999999

server is on the same network as the nas and able to ping and traceroute to it.

# ping -s tfnas9
PING tfnas9: 56 data bytes
64 bytes from tfnas9.xpedite.com (137.236.32.52): icmp_seq=0. time=0. ms
64 bytes from tfnas9.xpedite.com (137.236.32.52): icmp_seq=1. time=0. ms
64 bytes from tfnas9.xpedite.com (137.236.32.52): icmp_seq=2. time=0. ms
64 bytes from tfnas9.xpedite.com (137.236.32.52): icmp_seq=3. time=0. ms

# traceroute tfnas9
traceroute to tfnas9.xpedite.com (137.236.32.52), 30 hops max, 40 byte packets
1 tfnas9-mgmt.xpedite.com (10.205.88.94) 0.844 ms 0.379 ms 0.397 ms

daemon 1541 1 0 Oct 29 ? 0:00 /usr/lib/nfs/statd
root 1542 1 0 Oct 29 ? 0:00 /usr/lib/nfs/lockd
root 2227 1 0 Oct 29 ? 5379:20 /usr/lib/nfs/nfsd
root 2223 1 0 Oct 29 ? 1795:20 /usr/lib/nfs/mountd

# ps -ef |grep rpcbind
root 1508 1 1 Oct 29 ? 16425:53 /usr/sbin/rpcbind

nothing in /var/adm/messages relating to any errors, messages related to timeouts only show up when i try to write to the share..

the nas is also using version 3 of nfs.

/dbsarchive from tfnas9:/vol/dbsarchive
Flags: vers=3,proto=tcp,sec=sys,hard,intr,link,symlink,rsize=65536,wsize=32768,retrans=5,timeo=600
Attr cache: acregmin=3,acregmax=60,acdirmin=30,acdirmax=60

[YTC#1]

Ok, how about , from the client

/usr/bin/rpcinfo -u <server> nfs
/usr/bin/rpcinfo -u <server> mountd

[vpek...@gmail.com]

root@dbsarchive1 # /usr/bin/rpcinfo -u tfnas9 nfs
program 100003 version 2 ready and waiting
program 100003 version 3 ready and waiting
root@dbsarchive1 # /usr/bin/rpcinfo -u tfnas9 mountd
program 100005 version 1 ready and waiting
program 100005 version 2 ready and waiting
program 100005 version 3 ready and waiting

[Volker Borchert]

vpek...@gmail.com wrote:

> root 2227 1 0 Oct 29 ? 5379:20 /usr/lib/nfs/nfsd
> root 2223 1 0 Oct 29 ? 1795:20 /usr/lib/nfs/mountd

Up since Oct 2017, or longer?

These are NFS server processes. And I think they are consuming
far too much time.

> root 1508 1 1 Oct 29 ? 16425:53 /usr/sbin/rpcbind

This seems far too much time as well.

Do you use NIS (f.k.a. YP)? If so -
- For which services?
- Are the NIS processes running on client, server, and YP server?
- Does "ypcat" on those maps you use work and show the expected?

Do you use NISplus (a.k.a. NIS+)?
(Can't help with that, sorry.)

Is your DNS configured properly on both client and server,
can both resolve each other and themselves forward and reverse?

What does "rpcinfo" on client and server show?

At least with NFSv2, any problem with NIS, NIS+, DNS might cause
the NFS server to be unable to map client credentials to server
credentials, client access will happen as "nobody", and listing a
world r-x directory will work, but writing a world r-- file won't.

--

"I'm a doctor, not a mechanic." Dr Leonard McCoy <mc...@ncc1701.starfleet.fed>
"I'm a mechanic, not a doctor." Volker Borchert <v_bor...@despammed.com>

[YTC#1]

Next, as Volker had mentioned, check permissions. (and NIS/DNS)

And lets see what

nfsstat -m

shows


Also, does either the client or the server have IPMP configured ?

[vpek...@gmail.com]

There is no nis or nis + on this client nor is there ipmp configured. dns is properly configured and they can resolve each other.

root@dbsarchive1 # ls -ld .
drwxr-xr-x 4 mis prod 4096 Jun 27 13:12 .



root@dbsarchive1 # nfsstat -m
/dbsarchive from tfnas9:/vol/dbsarchive
Flags: vers=3,proto=tcp,sec=sys,hard,intr,link,symlink,rsize=1024,wsize=1024,retrans=5,timeo=15
Attr cache: acregmin=3,acregmax=60,acdirmin=30,acdirmax=60

root@dbsarchive1 # ps -ef |grep nfs
root 323 1 0 13:24:51 ? 0:00 /usr/lib/nfs/lockd
daemon 324 1 0 13:24:51 ? 0:00 /usr/lib/nfs/statd
root 17121 17106 0 10:13:58 pts/4 0:00 grep nfs
root@dbsarchive1 # ps -ef |grep rpcbind
root 288 1 0 13:24:50 ? 0:41 /usr/sbin/rpcbind
root 17123 17106 0 10:14:03 pts/4 0:00 grep rpcbind

[YTC#1]

Ok, you need to snoop (preferably from a 3rd party) when trying to write
and see if anything crops up there.

[Chris]

Just a long shot for 9, but where is the nfs server process in that
list ?...

Chris

[YTC#1]

On a client ??

[Chris]

Ok, I would be looking at the server end for a write fail, access
permissions primarily. And, for that particular client, is the
resource exported rw ?...

Chris