Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
NFS and changing network interface [Solved]
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Networking & Security
View previous topic :: View next topic  
Author Message
Nicias
Guru
Guru


Joined: 06 Dec 2005
Posts: 446

PostPosted: Sun May 31, 2015 4:53 pm    Post subject: NFS and changing network interface [Solved] Reply with quote

I have a laptop that I often use at my desk. When it is there I use a wired ethernet connection. When I am not at the desk, I use wifi.

I have it setup to turn off the wifi when the ethernet cord is plugged in. I manage this by adding a couple of lines in /etc/ifplugd/ifplugd.action
Code:
#!/bin/sh
# Copyright 1999-2004 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
#
# Gentoo-specific ifplugd.action
#
# This file gets called by ifplugd when it wants to bring an interface
# up or down.
#

if grep -q initng /proc/1/cmdline
then
    EXEC="/sbin/ngc"
    INITNG="yes"
else
    EXEC="/etc/init.d/net.$1"
    INITNG="no"
fi

case "$2" in
    up)
   if [ "${INITNG}" = "yes" ]
   then
       ARGS="-u net/$1"
   else
       /etc/init.d/net.wifi --quiet stop   # this line is new
       ARGS="--quiet start"
   fi
   ;;
    down)
   if [ "${INITNG}" = "yes" ]
   then
       ARGS="-d net/$1"
   else
       ARGS="--quiet stop"
       /etc/init.d/net.wifi --quiet start   # this line is new
   fi
   ;;
    *)
   echo "$0: wrong arguments" >&2
   echo "Call with <interface> <up|down>" >&2
   exit 1
   ;;
esac

export IN_BACKGROUND=true

if [ -x "${EXEC}" ]
then
    ${EXEC} ${ARGS}
    exit 0
else
    logger -t ifplugd.action "Error: Couldn't configure $1, no ${EXEC} !"
    exit 1
fi

However, some times changing network interfaces screws up my nfs mounts, and the only fix seems to be to restart.

I'm mounting the nfs mounts with this in my fstab
Code:

192.168.98.2:/export/media      /mnt/eridu/media        nfs      user,rw,_netdev,bg,auto,intr,soft,retry=10


Should I be doing this differently?


Last edited by Nicias on Wed Jun 03, 2015 10:15 pm; edited 1 time in total
Back to top
View user's profile Send private message
massimo
Veteran
Veteran


Joined: 22 Jun 2003
Posts: 1226

PostPosted: Mon Jun 01, 2015 7:13 am    Post subject: Reply with quote

Would AutoFS fit your needs?
_________________
Hello 911? How are you?
Back to top
View user's profile Send private message
charles17
Advocate
Advocate


Joined: 02 Mar 2008
Posts: 3664

PostPosted: Mon Jun 01, 2015 7:20 am    Post subject: Re: NFS and changing network interface Reply with quote

Nicias wrote:
Should I be doing this differently?
Maybe you might want to have a look at https://forums.gentoo.org/viewtopic-p-7611984.html#7611984 and https://wiki.gentoo.org/wiki/Network_dependent_services
Back to top
View user's profile Send private message
Nicias
Guru
Guru


Joined: 06 Dec 2005
Posts: 446

PostPosted: Mon Jun 01, 2015 9:07 pm    Post subject: Reply with quote

I don't think either of those work.

I suppose it boils down to what to do if the directory is in use. Mostly just a shell being open somewhere in the mount. If the connect interface changes, none of those solutions seem to fail gracefully. My best solution seems to be a restart.
Back to top
View user's profile Send private message
szatox
Advocate
Advocate


Joined: 27 Aug 2013
Posts: 3136

PostPosted: Tue Jun 02, 2015 7:56 pm    Post subject: Reply with quote

Something doesn't fit here.

Behaviour you report would be typical if it was NFS host that changes interface. Is it the case?
NFS goes dumb when server's IP changes after it's mounted. Even if the name reported is the same, there is IP specified in mount options and if that IP is unreachable connection fails and NFS client hangs.
To fix it you can ensure IP of NFS host doesn't vanish when you change connection. You can team or bridge those interfaces or just keep some script around that will assign static IP to your interface.
If it's client's interface that changes and hangs NFS it's really really weird, NFS server is almost stateless. Almost, because it does track mounts, but if a new machine requests data it could have mounted, server performs "internal" mount operation and carries on.

So, which setup is it?
Back to top
View user's profile Send private message
Nicias
Guru
Guru


Joined: 06 Dec 2005
Posts: 446

PostPosted: Tue Jun 02, 2015 9:28 pm    Post subject: Reply with quote

Host is always the same:
Code:

$ mount
....
192.168.98.2:/export/media on /mnt/eridu/media type nfs (rw,noexec,nosuid,nodev,bg,intr,soft,retry=10,vers=4,addr=192.168.98.2,clientaddr=192.168.98.6,user,_netdev)
$ ifconfig
lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 0  (Local Loopback)
        RX packets 71  bytes 3976 (3.8 KiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 71  bytes 3976 (3.8 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

wifi: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.98.16  netmask 255.255.255.0  broadcast 192.168.98.255
        inet6 fe80::1e4b:d6ff:fe7f:f7e9  prefixlen 64  scopeid 0x20<link>
        ether 1c:4b:d6:7f:f7:e9  txqueuelen 1000  (Ethernet)
        RX packets 26345  bytes 2753141 (2.6 MiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 43  bytes 5325 (5.2 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

wired: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.98.6  netmask 255.255.255.0  broadcast 192.168.98.255
        inet6 fe80::e2cb:4eff:fefd:710b  prefixlen 64  scopeid 0x20<link>
        ether e0:cb:4e:fd:71:0b  txqueuelen 1000  (Ethernet)
        RX packets 577993  bytes 565661534 (539.4 MiB)
        RX errors 0  dropped 11  overruns 0  frame 0
        TX packets 429500  bytes 64016889 (61.0 MiB)
        TX errors 0  dropped 0 overruns 0  carrier 1  collisions 0
$ cd /mnt/eridu/media
nrobbins@kish /mnt/eridu/media $ ls
<works fine>
<unplug ethernet>
<wait a moment>
$ ifconfig
lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 0  (Local Loopback)
        RX packets 71  bytes 3976 (3.8 KiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 71  bytes 3976 (3.8 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

wifi: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.98.16  netmask 255.255.255.0  broadcast 192.168.98.255
        inet6 fe80::1e4b:d6ff:fe7f:f7e9  prefixlen 64  scopeid 0x20<link>
        ether 1c:4b:d6:7f:f7:e9  txqueuelen 1000  (Ethernet)
        RX packets 26406  bytes 2758745 (2.6 MiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 46  bytes 5601 (5.4 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

wired: flags=4099<UP,BROADCAST,MULTICAST>  mtu 1500
        inet6 fe80::e2cb:4eff:fefd:710b  prefixlen 64  scopeid 0x20<link>
        ether e0:cb:4e:fd:71:0b  txqueuelen 1000  (Ethernet)
        RX packets 578135  bytes 565710547 (539.5 MiB)
        RX errors 0  dropped 11  overruns 0  frame 0
        TX packets 429603  bytes 64034440 (61.0 MiB)
        TX errors 0  dropped 0 overruns 0  carrier 2  collisions 0

nrobbins@kish /mnt/eridu/media $ mount
<hang>
<new shell>
$ mount
<...>
192.168.98.2:/export/media on /mnt/eridu/media type nfs (rw,noexec,nosuid,nodev,bg,intr,soft,retry=10,vers=4,addr=192.168.98.2,clientaddr=192.168.98.6,user,_netdev)
$ cd /mnt/eridu/media
<hang>
<new shell>
$ dmesg
[86922.464063] nfs: server 192.168.98.2 not responding, timed out
[86965.568092] nfs: server 192.168.98.2 not responding, timed out
[87023.920110] nfs: server 192.168.98.2 not responding, timed out

I can umount -f it, but then I can't mount it back.
Back to top
View user's profile Send private message
szatox
Advocate
Advocate


Joined: 27 Aug 2013
Posts: 3136

PostPosted: Wed Jun 03, 2015 5:07 pm    Post subject: Reply with quote

Quote:
(rw,noexec,nosuid,nodev,bg,intr,soft,retry=10,vers=4,addr=192.168.98.2,clientaddr=192.168.98.6,user,_netdev)
NFSv4, clientaddr in mount options... I guess that's the culprit.
What about fixing that IP on client side? NFS4 uses TCP so you might not be able to remove that option completly, but you can still "restore" old IP upon iface switch
Back to top
View user's profile Send private message
Nicias
Guru
Guru


Joined: 06 Dec 2005
Posts: 446

PostPosted: Wed Jun 03, 2015 10:15 pm    Post subject: Reply with quote

That fixed it. I think.

I first tried to get my router to give me the same IP to both interfaces. It didn't want to do that.

Then I worked out how to change the MAC address of my wifi card to match the ethernet MAC. The router does ip's by MAC's, and is too stupid to notice that who used to be connected over wifi is now connected over ethernet with the same MAC. It gives me the same address.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Networking & Security All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum