/usr/openv/netbackup/bin/bpclntcmd -pn

[Zailar]

Hello All,

Master server hostname = master

Media server hostname = svr

It's have the virtual ip

It's media server solaris 11 always error if run /usr/openv/volmgr/bin/tpautoconf -t "Failed opening the device mappings file".

And check with  /usr/openv/netbackup/bin/bpclntcmd -pn not fixed.

root@svr:~# /usr/openv/netbackup/bin/bpclntcmd -pn
expecting response from server master
10.147.111.14 *NULL* 10.147.111.14 44593

root@svr:~# /usr/openv/netbackup/bin/bpclntcmd -pn
expecting response from server master
10.147.111.15 *NULL* 10.147.111.15 58432

root@svr:~# /usr/openv/netbackup/bin/bpclntcmd -pn
expecting response from server master
10.147.111.18 *NULL* 10.147.111.18 34526

root@svr:~# /usr/openv/netbackup/bin/bpclntcmd -pn
expecting response from server master
10.147.111.13 *NULL* 10.147.111.13 65242

root@svr:~# /usr/openv/netbackup/bin/bpclntcmd -pn
expecting response from server master
10.147.111.12 *NULL* 10.147.111.12 32864

root@svr:~# /usr/openv/netbackup/bin/bpclntcmd -pn
expecting response from server master
10.147.111.17 *NULL* 10.147.111.17 43984

root@svr:~# /usr/openv/netbackup/bin/bpclntcmd -pn
expecting response from server master
svr *NULL* 10.147.111.19 41948

Any solution ?

Thanks and Regards

[Marianne]

What is actual IP address of master and media server?
Are you using hosts files for name resolution or DNS?

You say:

Media server hostname = svr 
It's have the virtual ip

Does this mean media server is clustered?

Have you installed and configured cluster nodes as individual media servers (as per NBU requirement)?

Show us output of these commands on the media server: 
bpclntcmd -self 
bpclntcmd -hn <master>
bpclntcmd -ip <master-ip>

On master:
bpclntcmd -hn <media-server>
bpclntcmd -ip <media-ip>

[Faysal]

Note the following communication step to find network issue.

1. Verify the master and media server ip address are resolve hostnames.

bpclntcmd -ip master_server_ipaddress
bpclntcmd -ip media_server_ipaddress

2. verify the master amd media server hostnames are resolve address.

bpclntcmd -hn master_server_name
bpclntcmd -hn media_server_name

3. verify what name the master server sees as each machine's peer name.

bpclntcmd -pn

4. Test the bpcd connection between each machine and the master server.

bptestbpcd -host master_server

5. Run the command for connectivity.

bptestbpcd -debug

bptestbpcd -client <MediaServer_name> -debug
bptestbpcd -client <Master_name> -debug -verbose

# netstat -an | grep 13724 ( if you then on the client server run the command)

Then to test the bpcd port locally on the client server from the command prompt, run this telnet test using the loopback interface-

telnet localhost 13782 or

telnet 127.0.0.1 13782

verify these port

bash-3.00# netstat -a | grep 13724
      *.13724              *.*                0      0 49152      0 LISTEN
bash-3.00# netstat -a | grep 13782
      *.13782              *.*                0      0 49152      0 LISTEN

bash-3.00# telnet localhost 1556 (or remote server)
 

and also please check your system firwall

[Zailar]

Hello Marianne,

In /etc/hosts master just only 

10.147.111.19  svr

and media server just using host files not use option DNS

this interface in media :

onm0: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.19 netmask ffffffe0 broadcast 10.147.111.31
        groupname onm0
onm0:1: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.11 netmask ffffffe0 broadcast 10.147.111.31
onm0:2: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.10 netmask ffffffe0 broadcast 10.147.111.31
onm0:3: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.14 netmask ffffffe0 broadcast 10.147.111.31
onm0:4: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.15 netmask ffffffe0 broadcast 10.147.111.31
onm0:5: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.18 netmask ffffffe0 broadcast 10.147.111.31
onm0:6: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.13 netmask ffffffe0 broadcast 10.147.111.31
onm0:7: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.12 netmask ffffffe0 broadcast 10.147.111.31
onm0:8: flags=108001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,IPMP,PHYSRUNNING> mtu 1500 index 2
        inet 10.147.111.17 netmask ffffffe0 broadcast 10.147.111.31
onm_1: flags=100001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,PHYSRUNNING> mtu 1500 index 3
        inet 0.0.0.0 netmask ff000000 broadcast 0.255.255.255

Thanks

[Zailar]

Media Server

root@svr # /usr/openv/netbackup/bin/bpclntcmd -self
yp_get_default_domain failed: (6) internal yp server or client error
NIS does not seem to be running: (1) args to yp function are bad
gethostname() returned: svr
host svr: svr at 10.147.111.19
aliases:     svr     10.147.111.19

getfqdn(svr) returned: svr.
root@svr # /usr/openv/netbackup/bin/bpclntcmd -hn master
host master: master at 10.128.1.37
aliases:     master     10.128.1.37

root@svr # /usr/openv/netbackup/bin/bpclntcmd -ip 10.128.1.37
host 10.128.1.37: master at 10.128.1.37
aliases:     master     10.128.1.37

Master Server

root@master # bpclntcmd -hn svr
host svr: svr at 10.147.111.19
aliases:     svr     10.147.111.19

root@master # bpclntcmd -ip 10.147.111.19
host 10.147.111.19: svr at 10.147.111.19
aliases:     svr     10.147.111.19

[Zailar]

It's need REQUIRED_INTERFACE in bp.conf

[Marianne]

You did not answer my question about cluster and NBU installation?

Are all of the virtual IPs on the media server linked/mapped to different hostnames?
It looks as if bpclntcmd -pn is going through all of the virtual IPs trying to find a policy name for the host on the master.

[Zailar]

Sorry Marianne,

The media server cluster with SF.

The issue was resolved after add REQUIRED_INTERFACE and bpclntcmd clear host cache.

We don't knows where the justification logs.

Because 3 days ago, I was do it the same treatment, but not success.

[mph999 · Accepted solution]

REQUIRED_INTERFACE should not really be used now, it has been 'replaced' by PREFERRED_NETWORK

Plenty of examples in this TN

http://www.symantec.com/docs/HOWTO60779

.. showing how it should be used.

[revarooo]

If its a clustered server add the CLUSTER_NAME in with the actual cluster name. It should then only use that interface

[Zailar]

Thanks for information Marthin,

Revarooo it's need CLUSTER_NAME ?.because we install one by one.

It's mean cluster1 install with SERVER = cluster1 and cluster2 install with SERVER = cluster2 and not use option cluster in netbackup install progress.

[Marianne]

Media server installation and configuration in a cluster must be done for each local node name.

The only type of clustered config for media servers must be 'app_cluster' on the master server.

As stated previously - each IP address must be linked to a different hostname (preferrably in hosts file on the node and in DNS.)
The hostname used for NBU should resolve to one IP address only and must be the SERVER and CLIENT_NAME used in bp.conf on the node.

PS: 
You never told us which NBU version?

[Boomchi_t3]

Media Server in cluster ... I think its unsupported after 7.x . check the HA guied and compatibility guide 

If your intention is to use Tape library with the Meida server in Cluster then its not a good idea .

However since you mentioned that tpautoconf if failing . can you check the entries in vm.conf file as well .

[Zailar]

Hi Marianne,

The NBU version 7.6.1.

All the cluster server,  I just install like the single server . In the interactive installation question about cluster, the answer no.

[Marianne]

Tape in clustered media server is 100% fine. NBU is not clustered, only app_cluster config on master. Configuration of app_cluster is covered in NBU HA Guide. Please take note of my post above that each IP must be linked to different hostname.