Unexpected NFS Auto-unmount?

[kayasaman]

At this point I'm trying the

Code: Select all

timeo=1200


option, though to renew the IP via dhcp doesn't take that long (2 minutes) but still.... perhaps it might work?

[sabayonino]

Where sabayon is installed ? Laptop ? Desktop ? ...

is connection Wired or Wireless ?

[kayasaman]

On laptop wireless

On desktop wired

[sabayonino]

On laptop wireless

On desktop wired

and both have the same problem ...

On your Desktop , can you remove NetworkManager service (not uninstall) and run dhcpcd only ?

Code: Select all

# rc-update del NetworkManager default

dhcpcd sould be enabled

Code: Select all

# rc-update add dhcpcd default

reboot and check if this solution works

If not , enable Networkmanager again

Code: Select all

# rc-update add NetworkManager default

Code: Select all

# /etc/init.d/NetworkManager start --nodeps

[kayasaman]

I will definitely try that, but then what of the laptop which is also on 24 hours a day and doesn't get turned off very often or put to sleep?

I mean I would still need NetworkManager in case I went "roaming".

In the meantime I think I've found something interesting:

Code: Select all

#!/bin/sh

# these are the args passed by NetworkManager
MY_DEV_NAME="$1"
MY_NM_ACTION="$2"

# restart network mounts service in any case, networking setup is changed
# FIXME, remove: just restart myself, do not restart
# services that depend on me
net_fs="afs,cifs,coda,davfs,fuse,fuse.sshfs,gfs,glusterfs,lustre,ncpfs,nfs,nfs4,ocfs2,shfs,smbfs"
if [ "${MY_NM_ACTION}" = "up" ]; then
   mount -at ${net_fs} &
   MY_PID=$!

   # give script 20 seconds then kill
   # it happens with unavailable servers
   sleep 20
   MY_PID_ALIVE=$(kill -0 ${MY_PID} &> /dev/null)
   # kill if alive
   if [ "${MY_PID_ALIVE}" = "0" ]; then
      kill ${MY_PID} || kill -9 ${MY_PID}
      umount -at ${net_fs} -l
   fi

else # down

   # Lazy umount all filesystems
   umount -at ${net_fs} -l

fi


taken from:

Code: Select all

/etc/NetworkManager/dispatcher.d
01-netmount-up-down.rc  10-openrc-status


Does the above mean that NetworkManager will just restart itself and not the services mentioned? - could that be the cause of my issue?

[sabayonino]

ooh. it could be.

could you try "wicd" instead NetworkManager ?

net-misc/wicd
Available versions: 1.7.2.4-r2 {{X ambiance +gtk ioctl libnotify mac4lin ncurses nls +pm-utils}}
Homepage: https://launchpad.net/wicd
Description: A lightweight wired and wireless network manager for Linux


remember : remove NetworkManager service before running wicd

You could try on your pc-desktop

[kayasaman]

could you try "wicd" instead NetworkManager ?

I could, though I've always experienced problems with "wicd" but that may have just been a Fedora bug

My favorite part though "wicd-cli" I mean come on that's what UNIX is all about

I'll let you know soon what happens.

[kayasaman]

It looks like wicd fixed the issue....

so far I haven't had any NFS dropouts meaning that the culpret was NetworkManager so either it's a bug or programmed behaviour but either way not expected behaviour.

[sabayonino]

It looks like wicd fixed the issue....

so far I haven't had any NFS dropouts meaning that the culpret was NetworkManager so either it's a bug or programmed behaviour but either way not expected behaviour.

Oh Well.

I suggest file a Bug and report your issue and link to this topic for more info

Maybe NM is buggy or some sets are bad

Thanks



[edit] please can you rename the topic title ? It isn't a kernel Fault ( I hope !)

[kayasaman]

Title fixed and also a bug report submitted!

Thanks for all the help