Skip to main content

In looking through the topics around communication with Clients in firewalled environments there is the discussion around using host files on those clients when no DNS resolution is possible.  In configuring a cvfwd.txt distribution file in the client group there doesn’t appear to be a way to also include the IP addresses of the Media Agents or Commserve.  There is the notion of the Network Proxy Appliance, however again we have a DNS dependency.  For small groups of clients it doesn’t make sense to deploy said appliance, rather hosts file seems more reasonable. 

If the cvfwd.txt file contains the names of the Commserve and Media Agents why not give us the ability to control the IP address used ( for NAT”d networks ), and just leverage the current IP’s of the Commserve and MA’s if appropriate ( for non-NAT’d networks )?

Looking for a reason why the hosts file is the chosen method here, rather than providing all that functionality right in the cvfwd.txt file?  

 

 

The client’s hostname is what CommVault uses for communication. So you can have shortname\FQDN\IP as the value and\or utilized Data Interface Pairs for specific backup networks.

 

Configuring a Dedicated Backup Network Using Data Interface Pairs (commvault.com)


Testing this out now, thank you for the pointer!


Looked to be the ticket to get the IP addresses to show up in the cvfwd.txt file on the client side.  Thank you.


Looked to be the ticket to get the IP addresses to show up in the cvfwd.txt file on the client side.  Thank you.

Yep, most people use data interface pairs as a DNS override, although its original intention was to allow multi-homed servers be able to choose which network interface use for communications when there were multiple valid routes.

There was an enhancement a couple of years ago that also allows it to work with wildcards, so its much easier to set them en-mass at a group level if required. That is in the docs too.


Wildcards are how I configured access since I only have a single IP on the MA’s that I needed to populate on the client side.  I don’t think I could handle a NAT though with this approach, which is more of a one-off and should be reviewed before trying to build something out.  Just thinking if I had control over what IP went into the cvFwd.txt file for the MA’s/Commserve would be handy.  An override so to speak so if the client wasn’t able to resolve the “real IP” of the MA/CS due to address space overlap I could populate that file with the client Networks NAT’d IP.  

 

This seems like an opportunity to overhaul the use of DIP’s and Network Topologies to enhance how smaller scale network flexibility can be accommodated in lieu of using a Network Proxy Appliance with DNS dependencies.


Wildcards are how I configured access since I only have a single IP on the MA’s that I needed to populate on the client side.  I don’t think I could handle a NAT though with this approach, which is more of a one-off and should be reviewed before trying to build something out.  Just thinking if I had control over what IP went into the cvFwd.txt file for the MA’s/Commserve would be handy.  An override so to speak so if the client wasn’t able to resolve the “real IP” of the MA/CS due to address space overlap I could populate that file with the client Networks NAT’d IP.  

 

This seems like an opportunity to overhaul the use of DIP’s and Network Topologies to enhance how smaller scale network flexibility can be accommodated in lieu of using a Network Proxy Appliance with DNS dependencies.

For NAT, you can also use the ‘port forward’ configuration - you don't have to adjust the ports but the hostname option gives you a way to override the destination hostname/port at an individual client/group level Here is what it looks like from the commcell console/advance network option.

I knew some MSPs using this approach that allowed them to handle large amount of individualized tenants.

 

 


Reply