asked on

AIX - rpc.mountd - not registerd

My server is an IBM-AIX 4.2 version opertating system.
Problem: rpc.mountd daemon never stayes active.
If I start the portmap daemon : stopsrc -g portmap
Then start rpc.mountd or nfs : startsrc -s rpc.mountd or startsrc -g nfs
The the rpc.mountd stay active, but when I start the portmap daemon, then rpc.mountd goes inactive.

How do I get rpc.mountd to stay active?

gheist

oslevel -r please...
there is no AIX 4.2

yodanyo

ASKER

the Aix version is 4.2.1.0

gheist

basically rpc.mountd and nfsd registers with portmapper when started, they do not work without it, or after restarting it.
You can always use rpcinfo -p to see working rpc services

yodanyo

ASKER

I have used rpcinfo -p ejs1 and show that mountd is not working. I know that portmap needs to run to allow the resest of the services to run. I my testing I found that is you stop portmap and start rpc.mountd, then mountd will show running in rpcinfo. This where I'm trying to determine where the problem lies. When I have portmap runing , then I start rpc.mountd, and run rpcinfo -p esj1 you will see portmap but not mountd running.

esj2{root}#lssrc -a
Subsystem Group PID Status
syslogd ras 6422 active
portmap portmap 6170 active
inetd tcpip 5406 active
biod nfs 7764 active
nfsd nfs 8296 active
rpc.statd nfs 8590 active
rpc.lockd nfs 9118 active
rpc.mountd nfs inoperative

In the errpt log I see rcp.mountd

esj2{root}#esj2{root}#errpt -a|more
----------------------------------------------------------
LABEL: SRC
IDENTIFIER: E18E984F

Date/Time: Sun Mar 6 19:49:03
Sequence Number: 6758
Machine Id: 007105534C00
Node Id: esj2_boot
Class: S
Type: PERM
Resource Name: SRC

Description
SOFTWARE PROGRAM ERROR

Probable Causes
APPLICATION PROGRAM

Failure Causes
SOFTWARE PROGRAM

Recommended Actions
PERFORM PROBLEM RECOVERY PROCEDURES

Detail Data
SYMPTOM CODE
0
SOFTWARE ERROR CODE
-9020
ERROR CODE
256
DETECTING MODULE
'srchevn.c'@line:'281'
FAILING MODULE
rpc.mountd
----------------------------------------------------------

Is the problem with the daemon rpc.mountd or portmap?
would you need to remove and redeliver nfs package or is it script file?

I coped rpc.mountd, portmap,inetd.conf from the working node1 to the troubled node2.
This is a cluster configuration (HACMP). I had to failover everything to node1 where portmap and rpc.mountd is work to fix the problem on node2.

This weekend I changed IP addresses, and everything is working except for rpc.mountd.
Where does the problem lies?

In inetd.conf, rc.tcpip, services, rc.nfs, and rpc.pcnfsd there are no reference to rpc.mountd in the files.
This is also true on the working node node1.

Where to go from here, could it be a simple permission problem?

gheist

It has lived to the end. It is to old to keep working, especially interoperate with new clients, It is not supported anymore.
For NFS storage you can have a common linux or BSD box with two software mirrored huge disks, and this will stay up much longer than your H-available cluster, and cost only a small fraction of your cluster.

gheist

Your error report shows that rpc.mountd crashed.
inetd.conf is irreleavant in this case.
AIX nfs services are hypersensitive to DNS lookups timing out.

yodanyo

ASKER

I resolved the problem by redelivering NFS/TCP software and relinking the librarys.
This fixed the problem.

Thanks for everyones input

gheist

and you did not tell you have problem with static linkage ...

ASKER CERTIFIED SOLUTION

modulo

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

yodanyo

ASKER

The problem has been fixed for rpc.mountd.
I don't see problems with static linkage....What do you need to check for?