E64 nfs

Thomas Zweifel

2023-01-16 14:13:05 UTC

Hallo Eisfair Team

Nach dem Update auf die aktuellen nfs-Pakete, musste ich feststellen
dass die Geschichte nicht mehr funktioniert....

Auf einem Testrechner habe ich eine entsprechende Konfiguration frisch
erstellt, leider geht an der Stelle schon etwas schief:

Server
Activate configuration now (y/n) [yes]?
creating idmapd.conf file ...
creating exports file ...
creating hosts.allow file ...
creating hosts.deny file ...
creating services file ...
* Stopping NFS daemon 'rpc.mountd' ...
[ OK ]
* Stopping NFS daemon 'rpc.idmapd' ...
[ OK ]
checkproc: Empty pid file /var/run/rpc.statd.pid for /usr/sbin/rpc.statd
* Starting NFS daemon 'rpc.idmapd' ...
[ OK ]
* Starting NFS daemon 'rpc.mountd' ...
[ OK ]
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
* Starting NFS daemon 'rpc.statd' ...
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
[ FAIL ]
* Starting NFS daemon 'rpc.nfsd' ...
[ FAIL ]
Press ENTER to continue

Client
Activate configuration now (y/n) [yes]?
creating services file ...
umount: /var/lib/nfs/rpc_pipefs: target is busy.
* Unloading NFS kernel module 'auth_rpcgss' ...
[ FAIL ]
* Loading NFS kernel module 'nfs' ...
[ OK ]
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
* Starting NFS daemon 'rpc.statd' ...
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
checkproc: Empty pid file /var/run/rpc.statd.pid for rpc.statd
[ FAIL ]
Press ENTER to continue

Nach dem händischen starten vom init.d/rpcbind, nfsserver und nfsclient
Skript lief es dann.

Herunterfahren geht dann auch in der Reihenfolge wie es bisher war:

service01test 3.2.7 # /etc/init.d/nfsclient stop
* Unmounting share '/home/tom/Testnfs' ...
[ OK ]
umount: /var/lib/nfs/rpc_pipefs: target is busy.
* Unloading NFS kernel module 'nfsv3' ...
[ OK ]
* Unloading NFS kernel module 'nfs' ...
[ OK ]
* Unloading NFS kernel module 'auth_rpcgss' ...
[ FAIL ]

service01test 3.2.7 # /etc/init.d/nfsserver stop
* Stopping NFS daemon 'rpc.mountd' ...
[ OK ]
* Stopping NFS daemon 'rpc.idmapd' ...
[ OK ]
* Stopping NFS daemon 'nfsd' ...
[ OK ]
* Stopping NFS daemon 'rpc.statd' ...
[ OK ]

service01test 3.2.7 # /etc/init.d/rpcbind stop
* Unloading NFS kernel module 'nfsd' ...
[ OK ]
* Unloading NFS kernel module 'auth_rpcgss' ...
[ OK ]
* Unloading NFS kernel module 'nfs_acl' ...
[ OK ]
* Stopping rpcbind ...
[ OK ]

Erneutes starten geht ebenfalls, auch wenn es unschön aussieht :-)

service01test 3.2.7 # /etc/init.d/rpcbind start
* Starting rpcbind ...
[ OK ]
* Loading NFS kernel module 'nfs' ...
[ OK ]
* Starting NFS daemon 'rpc.statd' ...
[ OK ]
* Mounting share '127.0.0.1:/Backup/Test' ...
mount.nfs: portmap query retrying: RPC: Program not registered
mount.nfs: portmap query failed: RPC: Program not registered
mount.nfs: portmap query retrying: RPC: Program not registered
mount.nfs: portmap query failed: RPC: Program not registered
mount.nfs: portmap query retrying: RPC: Program not registered
mount.nfs: portmap query failed: RPC: Program not registered
mount.nfs: requested NFS version or transport protocol is not supported
[ OK ]
* Loading NFS kernel module 'nfsd' ...
[ OK ]
* Starting NFS daemon 'rpc.idmapd' ...
[ FAIL ]
* Starting NFS daemon 'rpc.mountd' ...
[ OK ]
* Starting NFS daemon 'rpc.nfsd' ...
[ OK ]

service01test 3.2.7 # /etc/init.d/nfsserver start
mount: /proc/fs/nfsd: nfsd already mounted on /proc/fs/nfsd.
* Starting NFS daemon 'rpc.idmapd' ...
[ FAIL ]

service01test 3.2.7 # /etc/init.d/nfsclient start
* Mounting share '127.0.0.1:/Backup/Test' ...
mount.nfs: trying 127.0.0.1 prog 100003 vers 3 prot TCP port 2049
mount.nfs: trying 127.0.0.1 prog 100005 vers 3 prot UDP port 33849
[ OK ]

service01test 3.2.7 # df
Filesystem 1K-blocks Used Available Use% Mounted on
devtmpfs 16298852 8 16298844 1% /dev
tmpfs 16306764 0 16306764 0% /dev/shm
tmpfs 16306764 832 16305932 1% /run
/dev/md3 30801888 10551204 20152384 35% /
/dev/md1 42411 33403 5573 86% /boot
127.0.0.1:/Backup/Test 30801920 10550272 20153344 35% /home/tom/Testnfs

Für den Moment konnte ich die Panne also umgehen, indem ich ein
Startlink für rpcbind erstellt habe und die nfs Skripte wieder in der
Reihenfolge starte/beende wie es früher war.

Hoffe ihr schaut euch die Geschichte bei gelegenheit nochmals an (eilt
nicht). ;-)

Gruss Thomas