Symantec Net Backup Client Problem

carlosinfl · 04-21-2009, 03:14 PM

I have 6 servers in my server room. Three of them are RHEL 5.3 x64 and the remaining three I have are CentOS 5.3 x64.

Now I was able to install the Net Backup (Symantec) client on all six servers however it only is able to connect to the three that are RHEL 5.3. All the CentOS servers get connections errors. I checked to make sure that SELinux and the default Firewall are disabled. There is no iptables running or anything of that nature and they're all on the same subnet and network connecting on the same tcp/udp port.

I don't want to tell Symantec support that I am running CentOS because they will just say "Sorry, we don't support that". Does anyone know what could be the problem?

Code:

lt2fs3 ~]# cat /etc/redhat-release 
CentOS release 5.3 (Final)
[root@lt2fs3 ~]# uname -a
Linux lt2fs3.ideorlando.org 2.6.18-128.1.6.el5 #1 SMP Wed Apr 1 09:10:25 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux

Code:

lt2fs1 cwilliams]# cat /etc/redhat-release 
Red Hat Enterprise Linux Server release 5.3 (Tikanga)
[root@lt2fs1 cwilliams]# uname -a
Linux lt2fs1.ideorlando.org 2.6.18-128.1.6.el5 #1 SMP Tue Mar 24 12:05:57 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux

lazlow · 04-21-2009, 03:36 PM

I would post this over at the Centos forum @ Centos.org. This is pretty specific (and odd), so they are going to have the best shot at solving it.

Either way you will need to post the specific connection errors in order for anybody to be of any real assistance.

andymalato · 04-21-2009, 07:54 PM

It appears that your CentOS systems differ somewhat in their setup from your RHEL5 systems. Any indication in the logs on your Master server that would indicate a connection failure? In cases where a client wasn't able to connect, the most likely cause was TCP wrappers. You could try putting the following in /etc/hosts.allow with the ips of your media controllers

Code:

bpcd-stream: x.x.x.x
bpcd: x.x.x.x
vnetd: x.x.x.x

lazlow · 04-21-2009, 08:01 PM

I am beginning to wonder if the software itself does not look at cat /etc/redhat-release and act on what it finds there.

carlosinfl · 04-22-2009, 10:39 AM

OK all. It appears that changing /etc/redhat-release did not work. I thought it would but it does not. I was just missing the 'xinetd' package on my CentOS x64 machines and when I installed it and started the service, I saw it was listing again!

Code:

yum install xinetd
/etc/init.d/xinetd restart
chkconfig xinetd on
/usr/openv/netbackup/bin/bpcd -standalone
 netstat -a | grep -i "bpcd"
tcp        0      0 *:bpcd                      *:*                         LISTEN      
unix  2      [ ACC ]     STREAM     LISTENING     314142 /usr/openv/var/vnetd/bpcd.uds

Once I did all that - you can see the service is now listening on the correct port and the master server can now connect to my CentOS machine perfectly!