NFS confusion

[pctech321]

the host mirage1 is nfs master
and county is a client
see the configuration:

mirage1/root: / # lsnfsexp
/prod -root=county,access=county

on mirage1 i run

mirage1/root:/lsvg -l emcvg
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
----------skip-------------------------
lv06 jfs 154 308 2 open/syncd /prod

-----------skip------------

so on county1 i see that filesystem, its an nfs right so all writes goes to master

county/root: / > lsvg -l sanvg
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
-----------------skip----------------
lv06 jfs 46 46 1 open/syncd /prod
----------------skip-------------------

whats that it mean ...the /prod on county1 is showing LPs of 46 and it included in another volume
group here??, i understand that nfs should not be inclueded in client storage right ???

can any one clarify ...

thank you..

 

[khalidaaa]

It seems to me that you are overmounting the file system! You might be having the same file system defined on country under sanvg volume group.

To make sure that this is the case, try to unmount /prod and then cd to /prod again! i think you will be accessing the local file system by then that is defined on sanvg in country server!

Regards,
Khalid

 

[khalidaaa]

I've just recreated your problem on our test machines!

nfs master is node1 and nfs client is node2

on nfs client (node2):
# lsvg -l rootvg
rootvg:
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
testlv jfs2 3 3 1 open/syncd /test

# lsnfsexp
/test -sec=sys:krb5p:krb5i:krb5:dh:none,rw,access=node1

On nfs master (node1):

# lsvg -l rootvg
rootvg:
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
testlvs jfs2 4 4 1 closed/syncd /test

basiclly i created an lv on node1 and exported it to node2 and i mounted it from node2 which has already another lv with the same name underneath it! So that's the case with your nfs!

now when i unmount the /test on node2 and try to mount it manually again (as if it is local mount point) it will mount by it will be locally and it may contain content that is different from the ones in node1 because it is local!

I hope i'm not confusing you more! Is that clear?

Regards,
Khalid

 

[pctech321]

thank you so much khalid ..ur right
here i got another problem
on master host i have a dir /var/oracle needs to be exported to another clinet
node
on master i done:
mknfsexp -d /var/oracle -t rw -c county
even i make sure this by /etc/exports file
mirage1/root: /etc # more exports
/var/oracle -rw,access=county

on county i run smitty mknfsmnt
i provide the mount point as /tmp/app
and remainings of * feilds
and secutiry method feild is just sys

it runs and ends with status of fail with foll msgs

mount: 1831-011 access denied for mirage1:/var/oracle
mount: 1831-008 giving up on:
mirage1:/var/oracle
The file access permissions do not allow the specified action.

whats the problem ..

thank you..

 

[khalidaaa]

Try this on master:

/usr/sbin/mknfsexp -d '/var/oracle' '-B' -S 'sys' -t 'rw' -c 'county' -r 'county'

and then make sure you get such an output of exportfs on the master:

# exportfs
/var/oracle -sec=sys,rw,access=county,root=county

Then go to the client (county) and do this:

/usr/sbin/mknfsmnt -f '/tmp/app' -d '/var/oracle' -h 'mirage1' -M 'sys' '-N' '-a' -t 'rw' -w 'bg' '-Y' '-Z' '-X' '-H' '-j' '-q' '-g'

now it should work! i think your problem is with the sys permission on server!

I don't usually use straight command line for this. I always use the smitty nfs to do that so what you see above from options are taken from the smitty nfs command image

Good luck.

Regards,
Khalid

 

[pctech321]

guys does any one know , is there any option in the smitty mknfsmnt feildsthat brings system down ....today i tried to mount one of file system on the aix server which is at 4330-10 , after i filled up all * feilds it started running and at some point , server get shutdown itself ...

its weird ....any help ,

thank you ,

 

[khalidaaa]

Trying looking into the errpt -a | more for errors!

I doubts that it is something with the nfs that caused the downtime!

Regards,
Khalid