Backup Central

Hi everybody,

we have a problem with Postgres on linux doing backup with savepnpc.
OS is Redhat 5.7, networker version is 7.6.2.7.

Node is running a PostgresDB, backup goes with savepnpc. Pstclientsave
sends WAL-Logs from client to server. Sometimes (not very often) the
following messages appears in savepnpc.log:

3817:save: Using "clientname" as server
39076:save: RPC warning: RPC cannot make a procedure call: Create failed:
Program not registered

After that it continues with
3817:save: Using "networker server" as server
and backup goes on.

Savesets on the server are complete, so it seems the saveset is sent again
with the right servername and saved on tape as it should be. However,
script sends back a returncode "1" and our Postgres-admin asks why backup
has gone wrong.

Any ideas how to avoid this?

Regards, Christian


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER

This might be caused because of the servers file. As far I have experienced on linux a servers file is not created under /nsr/res folder. You will have to manually create a file by the name "servers" no extensions under /nsr/res on that client machine. Open servers file with a text editor eg. vi and put in the name of the NetWorker server. Once done restart he nsrexecd service on the client machine. That should get it going.

Rovin A. D'Souza
rovinabi < at > gmail.com

-----Original Message-----
From: EMC NetWorker discussion [mailto:NETWORKER < at > LISTSERV.TEMPLE.EDU] On Behalf Of =?utf-8?Q?Christian_Jurecka?=
Sent: Wednesday, April 18, 2012 7:41 PM
To: NETWORKER < at > LISTSERV.TEMPLE.EDU
Subject: [Networker] linux client trying to use itself as server

Hi everybody,

we have a problem with Postgres on linux doing backup with savepnpc.
OS is Redhat 5.7, networker version is 7.6.2.7.

Node is running a PostgresDB, backup goes with savepnpc. Pstclientsave sends WAL-Logs from client to server. Sometimes (not very often) the following messages appears in savepnpc.log:

3817:save: Using "clientname" as server
39076:save: RPC warning: RPC cannot make a procedure call: Create failed:
Program not registered

After that it continues with
3817:save: Using "networker server" as server and backup goes on.

Savesets on the server are complete, so it seems the saveset is sent again with the right servername and saved on tape as it should be. However, script sends back a returncode "1" and our Postgres-admin asks why backup has gone wrong.

Any ideas how to avoid this?

Regards, Christian




via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER

Of course, RPC capability is mandatory for NetWorker. "program not registered" usually means that there is a (temporary) network related or port problem. In this case, as the client can not contact the hostname, it will obviously try to find a NW server on `localhost` as last chance.

In regard to: [Networker] linux client trying to use itself as server,...:

we have a problem with Postgres on linux doing backup with savepnpc.
OS is Redhat 5.7, networker version is 7.6.2.7.

Node is running a PostgresDB, backup goes with savepnpc. Pstclientsave
sends WAL-Logs from client to server. Sometimes (not very often) the
following messages appears in savepnpc.log:

3817:save: Using "clientname" as server
39076:save: RPC warning: RPC cannot make a procedure call: Create failed:
Program not registered

This could be any number of things. The suggestion you've already
received to verify the contents of /nsr/res/servers is a good one,
especially if there are any hosts on your network that you don't trust,
as having an empty servers file has security implications.

Since it manifests as an intermittent network problem, it could
potentially be on the client, on the intervening network gear, or on
the server.

The Installation guide for the server software and the "Technical Guidance
for Upgrades to EMC NetWorker" documents on powerlink have a number of
tuning suggestions for each NetWorker server platform. If you haven't
applied those suggestions, I highly recommend you investigate them.
Intermittent network issues could certainly happen if the server is
running short of some network-related resource.

After that it continues with
3817:save: Using "networker server" as server
and backup goes on.

One additional thing you could try is to modify the NetWorker client
properties so that the server passes "-s servername" to each save. You
can do this via the GUI by filling in the *server's* name (we use FQDN,
but your environment may differ) in the "server network interface" on the
"Globals (1 of 2)" tab for the client properties.

This may just mask the underlying problem, though.

Savesets on the server are complete, so it seems the saveset is sent again
with the right servername and saved on tape as it should be.

Are "Client retries" set to some positive value in the properties for the
group containing that client?

Tim
--
Tim Mooney Tim.Mooney < at > ndsu.edu
Enterprise Computing & Infrastructure 701-231-1076 (Voice)
Room 242-J6, IACC Building 701-231-8541 (Fax)
North Dakota State University, Fargo, ND 58105-5164


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER

The /nsr/res/servers file is actually the list of those Networker
servers which are allowed to backup the client. If the file is empty,
then all Networker servers can backup the client. This is like the -s
flag for nsrexecd (check its man page, it also discusses the servers file).

On 04/18/2012 05:30 PM, Dsouza, Rovinabishek wrote:
This might be caused because of the servers file. As far I have experienced on linux a servers file is not created under /nsr/res folder. You will have to manually create a file by the name "servers" no extensions under /nsr/res on that client machine. Open servers file with a text editor eg. vi and put in the name of the NetWorker server. Once done restart he nsrexecd service on the client machine. That should get it going.

Rovin A. D'Souza
rovinabi < at > gmail.com

-----Original Message-----
From: EMC NetWorker discussion [mailto:NETWORKER < at > LISTSERV.TEMPLE.EDU] On Behalf Of =?utf-8?Q?Christian_Jurecka?=
Sent: Wednesday, April 18, 2012 7:41 PM
To: NETWORKER < at > LISTSERV.TEMPLE.EDU
Subject: [Networker] linux client trying to use itself as server

Hi everybody,

we have a problem with Postgres on linux doing backup with savepnpc.
OS is Redhat 5.7, networker version is 7.6.2.7.

Node is running a PostgresDB, backup goes with savepnpc. Pstclientsave sends WAL-Logs from client to server. Sometimes (not very often) the following messages appears in savepnpc.log:

3817:save: Using "clientname" as server
39076:save: RPC warning: RPC cannot make a procedure call: Create failed:
Program not registered

After that it continues with
3817:save: Using "networker server" as server and backup goes on.

Savesets on the server are complete, so it seems the saveset is sent again with the right servername and saved on tape as it should be. However, script sends back a returncode "1" and our Postgres-admin asks why backup has gone wrong.

Any ideas how to avoid this?

Regards, Christian




via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER


--

-- Yaron.


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER

You really need to show us the savepnpc command if you want us to
figure that out.

On 04/18/2012 05:10 PM, =?utf-8?Q?Christian_Jurecka?= wrote:
Hi everybody,

we have a problem with Postgres on linux doing backup with savepnpc.
OS is Redhat 5.7, networker version is 7.6.2.7.

Node is running a PostgresDB, backup goes with savepnpc. Pstclientsave
sends WAL-Logs from client to server. Sometimes (not very often) the
following messages appears in savepnpc.log:

3817:save: Using "clientname" as server
39076:save: RPC warning: RPC cannot make a procedure call: Create failed:
Program not registered

After that it continues with
3817:save: Using "networker server" as server
and backup goes on.

Savesets on the server are complete, so it seems the saveset is sent again
with the right servername and saved on tape as it should be. However,
script sends back a returncode "1" and our Postgres-admin asks why backup
has gone wrong.

Any ideas how to avoid this?

Regards, Christian


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER


--

-- Yaron.


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER

First of all,

thank you everybody for your suggestions.

As a first solution I added an additional servers file and I hope this
helps.
Client retries for this special group are set to "2", so this might be the
reason why save is started again, although I always thought client retries
are only used for not-ok-savesets initiated by the group and not by those
initiated by savepnpc on the client.

However, if the problem appears networker sends back a returncode that
isn´t 0. Our DB-Admin asks for this returncode in his script and asks me
why backup isn´t ok (although it is). Can there be found more information
about the networker return codes? I just can find "0" for ok and "1" for
not ok. Is there some additional documentation?

Regards,
Christian


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER