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

EMM server - wrong hostname

Created: 15 Mar 2013 | 15 comments

- Netbackup 6.5.4

- Briefly changed the server to a new DNS domain during a physical move, but reverted changes as that is not supported

- The EMM server is stuck with the "new" domain name, and thus I can't inventory the robot. When I do, I get "EMM Server: netbackup1.new.com" and all the fields are blank. I presume I need to change it back to netbackup1.old.com in order to see the device host/robot. Doing nbemmcmd -deletehost and -addhost doesn't work; "server" isn't a valid machinetype for deletehost, and trying to add netbackup1.old.com results in it telling me that it's already in the database. bpmedia shows that all the tapes are assigned to netbackup1.new.com even though it was only changed a week ago.

- Current backup policies are running fine and the robot/tape drives are responding ok. 

See below:

nbemmcmd -list -verbose:

The following hosts were found:

netbackup1.new.com
        MachineName = "netbackup1.new.com"
        MachineDescription = ""
        MachineNbuType = server (6)
netbackup1.old.com
        ClusterName = ""
        MachineName = "netbackup1.old.com"
        GlobalDriveSeed = "VEND:#.:PROD:#.:IDX"
        LocalDriveSeed = ""
        MachineDescription = ""
        MachineFlags = 0x66
        MachineNbuType = master (3)
        MachineState = active for tape and disk jobs (14)
        NetBackupVersion = 6.5.4.0 (654000)
        OperatingSystem = linux (16)
        ScanAbility = 5
 
bp.conf:
SERVER = netbackup1.old.com
CLIENT_NAME = netbackup1.old.com
EMMSERVER = netbackup1.old.com
VXDBMS_NB_DATA = /usr/openv/db/data
(etc etc)
 
vm.conf:
MM_SERVER_NAME = netbackup1.old.com
(etc etc)

Any ideas?

Operating Systems:

Comments 15 CommentsJump to latest comment

sazz.'s picture

Ahh.. the disaster already happened and as you see the database shows inconsistency. As we see that it is still showing the master server name rightly in vm.conf and bp.conf and the backups are running fine. My concern is have you tried checking your restores, are they working or you do see your restores?
Try to movemedia on couple of medias to old server and check they allow you to do the restore fine.

As you can't remove this entry call Symantec support and see if they can help to remove it from the database.

netbackup1.new.com
        MachineName = "netbackup1.new.com"
        MachineDescription = ""
        MachineNbuType = server (6)

Please make sure at the earliest that all is working fine in your environment as changing the master server name gets into the datbase and can get lot of issues and problems. This can't be done and needs consulting services.

The other best option if you still have issues. As you have already renamed the machine to old name uninstall the NBU completely from the server, install the NBU server again and do the catalog recovery prior to the point where you have changed the name of the server and then run the phase1 and phase2 import for the tapes after that till now.

jpet's picture

Thanks, and ugh. Last time I called them they said 6.5 was EOL and wouldn't support it after that last case, but I guess I'll try anyway. We're junking the whole system in about 6 months so it's not cost-effective to bother upgrading. I'll see if I can get them to quote me consulting services. I understand why they don't want to let us access the database.. but.. it'd be so simple.. just change one thing.. ah well.

I'm not a NBU expert (at least not enough to do the Other Best Option) so that would also cost money to have a consultant come in and do that. But I'll see what they can do for me.

Marianne's picture

Pity you did not ask BEFORE making the changes... 

If you tell us what exact steps were followed here:

Briefly changed the server to a new DNS domain during a physical move, but reverted changes....

we may be able to assist.

We will need to know what was done at OS level as well as NBU level.

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

jpet's picture

I changed the hostname (hostname netbackup1.new.com), edited ifcfg-eth0 to the new IP/gateway, edited resolv.conf to the new DNS servers, edited yp.conf to point to the NIS new IP, edited /etc/hosts with new IPs/FQDNs, and edited vm.conf and bp.conf to the new FQDN. I then had trouble logging into the Java console (it was very slow) and after searching online that's when I found out that you're not supposed to try to change the hostname. I undid each step in reverse order until I could log in again normally. The Scratch tapes that were in the drive continued to be used normally so it wasn't until I tried to put new tapes in the drive that I noticed there was anything wrong.

jpet's picture

Also I had to run vmglob -set_gdbhost to new.com to match the hostname of the server. I've set it back, and -get_gdbhost goes to old.com, but that might have been what put the inconsistency in the database.

StefanosM's picture

I think that adding to the hosts file a line with the IP for the server, the old FQDN and new FQDN will temporary solve your problem, if you cannot reinstall netbackup and restore the database.

Do you have active support with Symantec?
If yes, the only way to get support is to open a case asking for netbackup database check before the upgrade.
This trick may be you only way to solve the problem. And of course to prosed with the upgrade.

Mark_Solutions's picture

I am surprised that what you have done has cause so many issue to be honest

Is the "proper" master name at the top of the bp.conf file?

Have you rebooted / used bpclntcmd -clear_host_cache since you changed everything back?

I would have expected it to take more than what you have described to actually cause such issues and for it to self correct after you reverted the changes - was there anything else that you did (patch / repair the installation etc?)

Just feels like you have missed something somewhere that you have not changed back

Authorised Symantec Consultant

Don't forget to "Mark as Solution" if someones advice has solved your issue - and please bring back the Thumbs Up!!.

jpet's picture

Yeah I'm pretty surprised too. I'm running 6.5.4 so -clear_host_cache isn't a valid switch for bpclntcmd. It's been rebooted a few times. 

I forgot as well, I had to run vmglob -set_gdbhost to new.com to match the hostname of the server. I've set it back, and -get_gdbhost goes to old.com, but that might have been what put the inconsistency in the database.

Mark_Solutions's picture

So what does nbemmcmd -getemmserver say?

Wondering is -setemmserver or even -machinealias will help you?

The trouble with the vmglob is that things can get in a mess!

Authorised Symantec Consultant

Don't forget to "Mark as Solution" if someones advice has solved your issue - and please bring back the Thumbs Up!!.

jpet's picture

getemmserver says netbackup1.old.com. When I try to delete the alias netbackup1.new.com it doesn't give an error but it's still there.

Mark_Solutions's picture

It is reading it from somewhere then - check your conf files again to see if it is in one of them somewhere - all servers

Authorised Symantec Consultant

Don't forget to "Mark as Solution" if someones advice has solved your issue - and please bring back the Thumbs Up!!.

Dyneshia's picture

What does the following return : 

nbemmcmd -getaliases -machinealias -machinetype master -machinename netbackup1.new.com

nbemmcmd -getaliases -machinealias -machinetype master -machinename netbackup1.old.com

You could run a NBSU and search in the files where netbackup1.new.com is still referenced.

It may be necessary to get a nbdbunload as there may be a reference at the table level and can only be removed by a script from backline.  Unfortunately since you are on an unsupported version of Netbackup, you may need to contact your sales rep to get an exception.

jpet's picture

Both return the following:

NBEMMCMD, Version:6.5.4
The following aliases were found for the alias: "netbackup1.old.com"
netbackup1.old.com
netbackup1
netbackup1.new.com
Command completed successfully.
 
Also: 
nbemmcmd -machinealias -deletealias -alias netbackup1.new.com -machinetype server
NBEMMCMD, Version:6.5.4
The function returned the following failure status:
invalid host name (136)
Command did not complete successfully.
 
Mark_Solutions's picture

I think the command should be:

nbemmcmd -deletealias -alias netbackup1.new.com -machinename netbackup1.old.com

You may need to add -machinetype on the end - possibly as Master rather than Media

See if this helps

Authorised Symantec Consultant

Don't forget to "Mark as Solution" if someones advice has solved your issue - and please bring back the Thumbs Up!!.

rahul.arora123's picture

Hi ,

I am getting one error as below:

Unable to connect to EMM database 

error no:-77