This post is more than 5 years old

100 Posts

4387

January 29th, 2013 01:00

backup of Linux server in SELinux

Hi,

I have a Legato server - Win2008-R2 with Legato V 7.6.4.5 - build 1088 and a new legato client running RH Linux 6 with SELinux (enhanced security) activated. In the past, all my Linux clients were not in SElinux mode. A backup worked immediately after installation and configuration. My server is called LGT2 and the new Linux client is called RHRC. Now, i get the following error:

86494: Unable to start the index connection: Permission denied. 'root' on RHRC must have remote access privilege to client LGT2.

39077: save: Permission 'root' on RHRC must have remote access privilege to client LGT2.

In the client RHRC i have specified root and its `password` in the fields Access - Remote user and password.

I have to tell that a lower version of Legato was installed on RHRC by the kicktstart procedure.. I did an rpm -e and than a rpm -Uvh of the version V 7.6.4.5.

After the installation of the V 7.6.4.5 i was told to run 2 commands related  to SELinux. By i lost that information.

How to proceed now to get it to work ?

Rgds,

Johan

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

January 30th, 2013 10:00

This is typical error which we discussed once which happens when you do backup of the box (Linux) before so called final reboot (given that whole box baseline has been tweaked to final names).  You can recognize it via nsradmin against box where you will still see the response of the box to be localhost.  That in return, when communicating with server, causes this sort of error.

100 Posts

January 29th, 2013 01:00

William,

The client shortname is not in the loopback line.....

I have disabled the SELinux and I still have the problem..... so it must be something else

Rgds,

Johan

445 Posts

January 29th, 2013 01:00

Johans,

I think you may have a misconfiguration in the etc/hosts file: -

From Support solution on EMC Support Website ref: esg69990

Cause:

Client's /etc/hosts file contained:

127.0.0.1 localhost.localdomain (client_name)

Resolution:

Remove the client's shortname from the loopback entry in the /etc/hosts file.

Rename nsr/res/nsrladb on the client and restart Networker services on the client.

Regards,

Bill Mason

736 Posts

January 29th, 2013 01:00

Hi Johan,

You should also be aware that having SELinux activated is not supported with NetWorker 7.6.x, so you are likely to get inconsistent results.  This is supported with NetWorker version 8.

-Bobby

2 Intern

 • 

253 Posts

March 12th, 2013 07:00

I have the same problem but do not know how to see it in nsradmin.

Is the correct solution to remove and reinstall the client or can it be fixed otherwise?

Thanks

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

March 12th, 2013 08:00

echo p | nsradmin -p 390113 -i - -s

2 Intern

 • 

253 Posts

March 12th, 2013 10:00

Thank you, mine is messed up.

Do I have to delete the client and start over or can it be changed?

2 Intern

 • 

253 Posts

March 12th, 2013 11:00

I have deleted the client and reinstalled it but it did not solve the problem.

Do I need to uninstall the server?

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

March 12th, 2013 13:00

Larry Alexis wrote:

Thank you, mine is messed up.

Do I have to delete the client and start over or can it be changed?

I think you do not read with understanding - this issue is not related to NW client or anyhow to NW.  This is an issue where Linux installation is not completed yet or not done properly. 

2 Intern

 • 

253 Posts

March 13th, 2013 04:00

So, if I follow what you are saying, I will need to unistall the NW server software and reinstall it now that the server itself is "finalized".

I did install it before everything was completed on the server, just for testing, but it seems to have kept some of the settings. I appears that the problems are only with the hosts that were used in testing.

How can I ensure that it will not keep any of these when I reinstall it?

Will I have to reinstall NW on all the effected clients too?

Or am I totally off-base with this assumption?

The servers are now up and running and completely functional except for the restore command

2 Intern

 • 

253 Posts

March 13th, 2013 04:00

Thank you, it is making sense now.

I am runnung AIX 7.1 and will start looking for solutions there

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

March 13th, 2013 04:00

Nope, you can uninstall NW software as much as you won't, but it won't change a thing - information is given by OS and therefore backup agent is just using it.  Uninstalling it and installing it back won't make difference.

If the problem is that system is identifying itself as localhost, then first place to look for is /etc/hosts as certain Linux distros used to make such setup when rolled out.  If not, it is buried more deep into Linux configuration from which system gets its name and id which is then broadcasted to the world.

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

March 13th, 2013 08:00

I just configure backup server on AIX 7.1

No Events found!

Top