Actually, this is a known bug with NW on RHEL6. With RHEL6 they
switched from portmapper to rpcbind. The client has issues with
registering itself with rpcbind, so rpcinfo -p doesn't show the service
at all. I had an SR opened for that (42683384) and this is the reply I got:
Yaron
Apologies for the delay in getting back to you, I've been in discussions
with NetWorker core development engineering and now it's clear why
rpcinfo -p is not showing the NetWorker daemons as registered with
rpcbind on RHEL 6 Enterprise.
If you stop NetWorker on the RHEL 6 Enterprise server and run:
strace nsrexecd -D9
this shows that the attempt to bind to rpcbind is failing due to:
Cannot register service with portmapper: Authentication error; why =
Client credential too weak
If you create a file in /etc/sysconfig called rpcbind with the following
content:
RPCBIND_ARGS=-i
This will start rpcbind in insecure mode when it is restarted, although
I suspect this is not the best idea for security reasons.
From the man page for rpcbind:
-i ``Insecure'' mode. Allow calls to SET and UNSET from any
host.
Normally rpcbind accepts these requests only from the loopback
interface for security reasons. This change is necessary for
programs that were compiled with earlier versions of the rpc
library and do not make those requests using the loopback
inter-
face.
If you start rpcbind in insecure mode and then run rpcinfo -p then the
NetWorker daemons do show up in the rpcinfo output on systems running
RHEL 6 Enterprise.
NetWorker is compiled with the earlier version of the rpc libraries to
allow compatibility with older versions of Linux and Solaris which do
show nsrexecd with rpcinfo -p correctly.
As the reporting of the NetWorker programs with rpcinfo -p is a
non-critical issue (as it does not impact the ability of NetWorker to
correctly backup and recover data) the issue will not be addressed until
the versions of Linux and Solaris using the older version of the rpc
libraries are no longer supported.
Please let me know if you require anything further on this service
request or if it can be closed.
Many thanks
Regards
Mark
On 06/01/2013 12:22 AM, Tim Mooney wrote:
In regard to: [Networker] Linux RHEL 6.4 client issue, Matthew Powell
said...:
Anyone have a Linux RHEL 6.4 client that could push a save to the server
which is Networker version 8.0 on Solaris 10 but the backup server not
be able to get to the client from a save job. Get RPC error, program not
registered. Firewall is off and still no luck. Any suggestions would be
most appreciated. Thanks and have a great day.
Joe already mentioned that selinux should be disabled on the RHEL client,
or the RHEL client admin needs to make the (unsupported) changes to
the selinux attributes on the client to get it to play nicely. Those
have been posted to this mailing list in the past.
Also, if the RHEL 6.x client was done as a minimal install, it probably
doesn't have rpcbind installed. nsrexecd has some built-in
portmapper/rpcbind capability, but I've always found that it works
better if portmapper or rpcbind is running on the clients. Have the
client admin check that rpcbind is installed, enabled for auto-start,
and running, and then have him or her restart nsrexecd.
You may also want to run
rpcinfo -p the-client-name
from your NetWorker server, to see if it can see nsrexec or program #
390113.
Tim
--
-- Yaron.
|