cancel
Showing results for 
Search instead for 
Did you mean: 

Netbackup 4.5 client gives connection time out

List_Man
Level 4
Hi all,

In our configuration we try to add linux and Windows2000 server clients. Starting the backup or retrieving the host properties from the master server says the network connection times out, as well as in the defined policies for the client.

What can be the case? Increasing CLIENT_CONNECT_TIMEOUT? A license?

Firewall settings are open to let pass traffic through between master server and clients. bpclntcmd give OK output both on the server as the clients.

Thanks, Listman
11 REPLIES 11

MayurS
Level 6
Update the physical hostname to IP mapping on client as well as server
As in windows update the Drivers/Etc/Hosts file.Message was edited by:
Mayur Shinde

List_Man
Level 4
We did so already. Pinging and nbclntcmd work fine.

Just increased the CLIENT_CONNECT_TIMEOUT to 900 on the server, but that didn't work either.

What else can there be wrong?

MayurS
Level 6
Terminate the" tracker.exe" on both the master and the client.

Refer to this document for details
http://seer.support.veritas.com/docs/267253.htm

MayurS
Level 6
Hey also refer to this document it also says the same thing as my above post.
http://seer.support.veritas.com/docs/275105.htm

Below is an excerpt from the document
Given below are common reasons for Status 41 error codes:
i.e. Network Connection timeout.
1. Network issues
2. Fragmented Drives
3. Corrupt file
4. Third-party software conflict (AntiVirus)
5. Client file system utilizing too many resources
6. If the client is utilizing 3.4.x, VERITAS recommends upgrading to 4.5 Feature Pack 6 (FP6) or later

Steps to try:

* Disable VERITAS Snapshot Provider (VSP) / Open Transaction Manager (OTM) on backup
* Disable AntiVirus
* Disable Tracker.exe

MayurS
Level 6
WHat else can be there?
Try these
1) Verify that the client version is equal to the master version.
2) try tracert and see whether you are reaching the client through a desired path. and vice-versa too.Message was edited by:
Mayur Shinde

List_Man
Level 4
Just did a restart of tracker on the W2k client. It didn't help.

bplclntcmd commands show there is a link and the command says it can find clientes and servers.

The exit status for the backup jobs changed from 41 to 54 after having set CLIENT_CONNECT_TIMEOUT to 900 on the server.

The versions on all machines are Netbackup 4.5 MP6

The Master Server is Solaris 9. The clients having problems are Linux and W2K Server. CLients with Solaris work fine.

Thanks again for any hints.

Listman

Poon_William
Level 4
Hi,

Perform a network transfer test by ftp files between master server and the client with hash on. The transfer should be smooth without pausing. If the data transfer is slow and pause in between, then problem is on your network setting.

Check your network setting on Master, Media server, clients and network switch ports. All must be at a fixed speed (100 or 1000) and full duplex. Auto-Negotiate is bad for any network activities. This is the most common cause for Network connection timeout.

WP

List_Man
Level 4
Network connection is OK. A copy with scp (secure copy) of a 47 MB file costs 10 seconds. That shouldn't be it.

Any hints on port usage by Netbackup?

MayurS
Level 6
Hi,

This will happen if there is issue in the communication or KEEP_ALIVE packets between the host and the master.

Have you tried restarting the netbackup services and checking the network card properties of the client. THe NIC should be sent to FULL DUPLEX.Message was edited by:
Mayur Shinde

Allan_Perocho
Level 3
hi,

try this one, i read this from veritas news group

1. Shut down the Vertias client service
2. Disable the NIC on the backup side
3. Look for the programs bpbkar32.exe , bpfis.exe, and
bpjava-msvc.exe . If they are running, end them.
4. Re-enable the NIC
5. Start the Vertias client service againMessage was edited by:
Allan Perocho

List_Man
Level 4
Hi All,

It is all resolved. The solution was easier than I thought: Just add the hostname in the hosts file of the system. For windows:

C:\Windows\System32\Drivers\etc\hosts

putting the host address in lmhosts won't work. For Unix:

/etc/hosts

Netbackup is doing a reverse lookup when establishing a connection. To check the bpcd process and network connection , give a:

telnet bpcd

The correct output should be:

Trying 10.xx.xx.xx...
Connected to
Escape character is '^]'.

Then give a return and the connection is killed. The connection should work either way: from the client as well as from the Netbackup server.
If there is no reponse from the client there will be only the '
Trying 10.xx.xx.xx...' message until telnet dies with the message: 'telnet: Unable to connect to remote host: Connection timed out'

Netbackup stores log information in files of the different services it has. In this case on the Windows Client:

\logs\bpcd

You may have to create the directory with the script that comes with installation of the windows client: \logs\mklogdir.bat
On Linux/ Unix you'll have to create the log directories by hand in /usr/openv/netbackup/logs/bpcd. If these directories are not there Netbackup won't store message, as I saw it.

The ports used by Netbackup should be let through by a firewall, if there is one in your network or on your local machine

Thanks you for all the other advices, that helped me discard any other sources of problems.