NFS not working - "writing fds to kernel failed"

Message

dsegel · Post by **dsegel** » Mon Jun 11, 2007 12:18 am

I get these errors when I start NFS on a new AMD64 X2 system I'm building:

Jun 10 17:13:24 [rpc.statd] Version 1.0.11 Starting
Jun 10 17:13:24 [nfsd] nfssvc: writing fds to kernel failed: errno 0 (Success)

I haven't been able to find out anything useful by searching for the error, and I'm a bit confused by the wording in any case (Success?).

When I try to mount the shared directories I get a "Connection refused" error.

My /etc/exports file looks like this:

Code: Select all

/home/share/ 192.168.1.0/24(rw,sync,no_root_squash,no_subtree_check)

Any tips on what I should try next?

vaxbrat · Post by **vaxbrat** » Mon Jun 11, 2007 12:37 am

What does this show?

Code: Select all

rc-status

Are you running a firewall on this server?

didymos · Post by **didymos** » Mon Jun 11, 2007 12:56 am

What kernel version do you have? Also, have you tried updating nfs-utils?

dsegel · Post by **dsegel** » Mon Jun 11, 2007 2:34 am

rc-status output:

Code: Select all

Runlevel: default
 gkrellmd                                                                                   [ started  ]
 hdparm                                                                                     [ started  ]
 lm_sensors                                                                                 [ started  ]
 local                                                                                      [ started  ]
 metalog                                                                                    [ started  ]
 net.eth0                                                                                   [ started  ]
 netmount                                                                                   [ started  ]
 nfs                                                                                        [ started  ]
 ntp-client                                                                                 [ started  ]
 ntpd                                                                                       [ started  ]
 postfix                                                                                    [ started  ]
 samba                                                                                      [ started  ]
 smartd                                                                                     [ started  ]
 sshd                                                                                       [ started  ]
 uptimed                                                                                    [ started  ]
 vixie-cron                                                                                 [ started  ]

NFS starts up, puts the original error message I posted into the system logs, and then just sits there. The nfsd processes (8 of them) are still running, but not accepting any connections.

Kernel is:

2.6.19-gentoo-r5 #4 SMP Sun Jun 10 16:06:29 PDT 2007 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 4600+ AuthenticAMD GNU/Linux

No firewall on this machine or between it and the client.

nfs-utils and nfsd are the latest stable versions in portage. I may try ~x86 next.

didymos · Post by **didymos** » Mon Jun 11, 2007 3:03 am

The reason I asked on the version is that rpc.statd start-up message you posted had 1.0.11 as the version number, but the only ebuilds with double-digit minor versions in portage are for 1.0.10 and 1.0.12.

dsegel · Post by **dsegel** » Mon Jun 11, 2007 4:01 am

It seems to be that way for the current (1.0.12) version. I just downgraded to 1.0.10 and it has the same problem. I also tried 1.1.0 (unstable) and it didn't work either.

I have another Gentoo system right next to this one and nfs runs fine on it. I'm beginning to think it has something to do with 64-bit, but I haven't explored that enough to know what to look for.

dsegel · Post by **dsegel** » Tue Jun 12, 2007 12:33 am

A short update: I still don't know what's causing the original NFS error, but it's not what was preventing the client system from mounting the exported filesystems - that turned out to be a DNS error.

qriff · Post by **qriff** » Fri Jun 15, 2007 12:59 pm

*bump*

Having same problem, things work but getting syslog errors.

Code: Select all

Jun 15 15:35:02 nfsd[24526]: nfssvc: writing fds to kernel failed: errno 0 (Success)
Jun 15 15:35:02 nfsd[24526]: nfssvc: writing fds to kernel failed: errno 0 (Success)

Chris_Hird · Post by **Chris_Hird** » Tue Jun 19, 2007 1:58 pm

dsegel wrote:rc-status output:

Code: Select all

Runlevel: default
 gkrellmd                                                                                   [ started  ]
 hdparm                                                                                     [ started  ]
 lm_sensors                                                                                 [ started  ]
 local                                                                                      [ started  ]
 metalog                                                                                    [ started  ]
 net.eth0                                                                                   [ started  ]
 netmount                                                                                   [ started  ]
 nfs                                                                                        [ started  ]
 ntp-client                                                                                 [ started  ]
 ntpd                                                                                       [ started  ]
 postfix                                                                                    [ started  ]
 samba                                                                                      [ started  ]
 smartd                                                                                     [ started  ]
 sshd                                                                                       [ started  ]
 uptimed                                                                                    [ started  ]
 vixie-cron                                                                                 [ started  ]

NFS starts up, puts the original error message I posted into the system logs, and then just sits there. The nfsd processes (8 of them) are still running, but not accepting any connections.

Kernel is:

2.6.19-gentoo-r5 #4 SMP Sun Jun 10 16:06:29 PDT 2007 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 4600+ AuthenticAMD GNU/Linux

No firewall on this machine or between it and the client.

nfs-utils and nfsd are the latest stable versions in portage. I may try ~x86 next.

Looks like portmap is not running? I think this could be the problem.

Chris...

Chris_Hird · Post by **Chris_Hird** » Tue Jun 19, 2007 7:17 pm

I have the portmapper running but still have the same problem as the above?

I did find this post on another forum and wondered if we have a similar problem I am running 2.6.20-r8
http://www.nabble.com/Bug-416066:-nfss ... 9220.html
Chris...

mlebek · Post by **mlebek** » Thu Jun 28, 2007 8:18 am

Try this /etc/export entry, is this working?:

Code: Select all

/home/share/ (rw,sync,no_root_squash,no_subtree_check)

col · Post by **col** » Mon Jul 02, 2007 4:43 am

I fixed this by removing kerberos support in nfs-utils & re-installing nfs-utils without it.

snIP3r · Post by **snIP3r** » Mon Jul 09, 2007 10:38 am

col wrote:I fixed this by removing kerberos support in nfs-utils & re-installing nfs-utils without it.

hi!

i have the same problem. but also compiling nfs-utils without kerberos support didnt fix it for me...

snIP3r

Alakhai · Post by **Alakhai** » Tue Jul 24, 2007 11:42 pm

I have the same problem trying to mount from a subnet to a "uppernet"
i explain i have 23.254* ip and a router that has a consecutive ip, it has a subnet 192.168.0.0/24
with my laptop i try to mount nfs on the 23 etc ip but it answer me the same of your...
in export and hosts.allow i set router ip and subnet
from same laptop the same day i successfully mounted a nfs from a same subnet

edit: i just tested both pc on same network fixing the proper config file, but i recive the same error

NFS not working - "writing fds to kernel failed"

NFS not working - "writing fds to kernel failed"

running services