Video Screencast Help
Symantec to Separate Into Two Focused, Industry-Leading Technology Companies. Learn more.

NBU client connectivity problems

Created: 21 May 2013 | 8 comments

In a VCS configuration, we have physical interfaces/IPs and service group logical interfaces/IPs.  When testing client (6.5.x and 7.x) connectivity from master server (first entry in bp.conf) using bptestbpcd, initial connections to the logical IPs fails; however, if we then run a connectivity test to the physical interface it works, and all subsequent tests to the logicals works just fine.  It's as if the connection to the physical "primes" the server for successful logical interface connections.

 

After a given amount of time (we haven't established exactly how long yet), this condition resets itself and we must "prime" again....leading me to believe there is a timeout associated with this.  Needless to say, when our backup jobs kick off for the logical interfaces, it fails too.

Operating Systems:

Comments 8 CommentsJump to latest comment

huanglao2002's picture

1 Please check the /etc/hosts/ or dns about physical/float  addressresolve

2 If you do a failover test on the resource ,you can run the "bpclntcmd -clear_host_cache" command on the both side.

3 use bptestbpcd -client 'hostname" from the server to test client connection.

mryanjax's picture

No problems with name resolution that I have found. It's DNS and both lookup and reverse lookup return correct address. Ping check to FQDN is successful as well.

The bptestbpcd -client command is what we're using to test connectivity and to display this problem we are seeing.

huanglao2002's picture

Can you post the test output?

from the server

telnet "client name" 1556

telnet "client name" 13724

 

from the client

telnet "server name" 1556

telnet "server name" 13724

mryanjax's picture

None of these connections connect; however, we run our bpcd -port 13782. Telnet acts the same way. The connection on port 13782 for the logical did not work for telnet until I connected to the physical on telnet.

This has to be some kind of configuration in xinetd where both telnet and bpcd run wouldn't it?

huanglao2002's picture

AS my experence,if encount network connect failed. i will do the below test.

1 from the local host

telnet localhost "port"

telnet `hostname` "port"

2 from remote host

telnet "remote host" "port"

telnet "ok remote host" "port"

After 1 and 2 test, i should know the  problem encounter on localhost /or network/remote host

Hope it help.

 

mryanjax's picture

The problem sounds close to this article:

http://www.symantec.com/business/support/index?pag...

But, the solutions do not seem to work...

LucSkywalker1957's picture

bptestbpcd -M backupserver -debug -verbose -client clientname

Initial handshake from backup server is on port 13782.

The NBU client does a reverse lookup and does a connect back to backup server with IP address on port 13724.

Port 1556 is needed for PBX and vnetd.

Make sure that those ports are opened bi-directionally (two-way traffic).

Marianne's picture

Is there any reason why your client is still on 6.5?

Connections on 6.5 is attempted on 13724 (vnetd) while 7.x will first attempt connection on 1556 (pbx), if that fails will try 13724. See http://www.symantec.com/docs/TECH136090

To answer your question about the timeout - it is NBU host cache that is refreshed every 1 hour. See http://www.symantec.com/docs/TECH136792

To troubleshoot connection failure, please ensure that bpcd log folder exists on both cluster nodes.
Run bptestbpcd with -verbose -debug options, then post output of command as well as bpcd log on active cluster node.

Supporting Storage Foundation and VCS on Unix and Windows as well as NetBackup on Unix and Windows
Handy NBU Links