cancel
Showing results for 
Search instead for 
Did you mean: 

Communication errors..

NBU35
Level 6

Master : SOlaris 10, NBU 7.6.0.3
CLient : NetBackup-Solaris10 7.0
Media: NetBackup-HP-UX11.31 7.6.0.3

Master is able to communicate with client, howvere there is some issue in connectivity of media server and client.
bptestbpcd is failing when it is executed from media server, also backups are failing with EC21.

Unable to find out any thing.
ping, telnet to bpcd, vnetd, pbx is working vice versa.
bpclntcmd is also provindg appropriate output.
I even have treid changing connect options from client attributes, but nothing has worked so far.

earlier this was working.

output of bptestbpcd attached.

Please suggest.

2 ACCEPTED SOLUTIONS

Accepted Solutions

Marianne
Level 6
Partner    VIP    Accredited Certified

We don't know what is causing the issue at this point in time.

Is bpcd updated when you bptestbpcd from the master?

Upgrade will ensure that all comms is happening on pbx (1556).

What exactly did you change as far as connect options are concerned and where?

View solution in original post

Marianne
Level 6
Partner    VIP    Accredited Certified
There is always a workaround or permanent fix for insufficient space in /tmp. 1) Creat TEMPDIR env variable that points to another location. 2) Read up on Solaris/Oracle site about creating a swap file and adding it to swap/tmp.

View solution in original post

18 REPLIES 18

Marianne
Level 6
Partner    VIP    Accredited Certified

Have you created bpcd log folder on the client?

You need this log on the client to troubleshoot.

Curious to know why client is still on 7.0?
This version will only listen on vnetd.
As from 7.0.1, all comms will attempt to use pbx (1556).

NBU35
Level 6

I beleive , it was working fine, so nobody noticed it. 

day before yesterday it was working fine/

Marianne
Level 6
Partner    VIP    Accredited Certified

You need bpcd log on the client....

Can you see in netstat -a output on the client if bpcd and vnetd are LISTENING?

There might be something wrong at TCP/IP level on the client that will most likely be fixed by a reboot.

You may as well upgrade the client at this point. 
 

 

Nicolai
Moderator
Moderator
Partner    VIP   

Use TCPDUMP (UNIX) or wireshark (windows). Thease tools are really cool for debugging flaky network connections.

Try to run to run and set :

bpclntcmd: -reverse_name_lookup prohibited

This will disable Netbackup reserve look-up featurte.

NBU35
Level 6

when I execute bptestbpcd from media server, it does not generate bpcd logs on client.

I can see vnetd and bpcd listening in output of netstat -a .

will upgrade fix the issue?

Marianne
Level 6
Partner    VIP    Accredited Certified

We don't know what is causing the issue at this point in time.

Is bpcd updated when you bptestbpcd from the master?

Upgrade will ensure that all comms is happening on pbx (1556).

What exactly did you change as far as connect options are concerned and where?

NBU35
Level 6

I changed the settings as follows, as in most of cases it has worked for me.

Random Port, Reserved Port, Deamon port only.

 

now when I try to install 7.6.0.1 on client, it is failing with following error:

 

Installing PBX...
Please wait while installation is in progress...
Installation completed Successfully
Installation log located here: /var/tmp/installpbx-1956-081015112110.log


ERROR:   There are 364872 kbytes of space in /tmp which
         is not enough to complete the installation.  About
         1122517 kbytes of space is required.
         Aborting...


ERROR:   A failure was detected running
         /tmp/mnt_NBU/netbackup/netbackup.7.6/client/NBClients/anb/Clients/usr/op
         env/netbackup/client/Solaris/Solaris10/cp_to_client master client /tmp/insta
         ll_trace.1661

 

Problem here is we cannot extend /tm as it is on physical slices and no space to extend.

Please suggest any other way to install NBU client without extending /tmp.

Nicolai
Moderator
Moderator
Partner    VIP   

Install from NFS share where the installation media is being made avaiable (The unpacked version of NB_CLT_7.6* files) or hack the install_client_files script.

 

Marianne
Level 6
Partner    VIP    Accredited Certified
There is always a workaround or permanent fix for insufficient space in /tmp. 1) Creat TEMPDIR env variable that points to another location. 2) Read up on Solaris/Oracle site about creating a swap file and adding it to swap/tmp.

NBU35
Level 6

Thanks Marianne, 

 

Issue has been fixed by following.

Set NB_TMPDIR

export NB_TMPDIR=required_path

then ./install script

installation finished successfully.

Marianne
Level 6
Partner    VIP    Accredited Certified

Backups and comms fine after the upgrade?

Good to hear!

Nicolai
Moderator
Moderator
Partner    VIP   

is the NB_TMPDIR a feature exclusive to Solaris on Netbackup or does it work for AIX,HP-UX, Linux as well ?

did not find anything in the manuals

Marianne
Level 6
Partner    VIP    Accredited Certified
I remember seeing something in a TN some time ago. Will see if I can find it.

sdo
Moderator
Moderator
Partner    VIP    Certified

This TN appears to indicate that it is Solaris only:

https://support.symantec.com/en_US/article.TECH189085.html

 

CRZ
Level 6
Employee Accredited Certified

Setting NB_TMPDIR is supposed to work in any UNIX/Linux environment and supersede the $tmpdir variable for the location of "temporary files directory" from 7.5 onwards. 

It looks like we meant to make TECH189085 a little more general (the issue was first reported on Solaris but could apply to any UNIX/Linux) but never got around to finishing that document update.

Marianne
Level 6
Partner    VIP    Accredited Certified
Chris, don't you think that this should be included in the Installation Guide?

Nicolai
Moderator
Moderator
Partner    VIP   

YES IT SHOULD :D

CRZ
Level 6
Employee Accredited Certified

¯\_(ツ)_/¯