Skip to main content
Question

Network Gateway Proxies

  • 15 July 2024
  • 3 replies
  • 35 views

Hi,

I have configured the following Topology which has all clients backing up to a HSX Cluster via Proxies.

 

The Proxies have the following installed. 

 

All backups & restore are going through the proxies and I am happy with this setup but I have two queries.

 

  1. Is there a way to tell what proxy the backup job going through ? You can’t see the proxy mentioned on the job logs or under streams running the job.  You also can’t see if when you add the Network Gateway column to your job controller window. 
  2. When testing restores I stop the CV services on two of my three proxies. I run a restore job and the job fails and it says it can’t communicate with a proxy “name” - my thinking here is it tries to use the same proxy for restore as it did for backup and as its off its fails. If I select Advanced Options → Use MediaAgent and select the only running proxy then the restore job completes. Is there a way for the restore job to look for a Proxy that's able to communicate rather than me having to tell the restore job ?

Thanks 

3 replies

Damian Andre
Rank
Userlevel 7
Badge +23
  1. I am assuming you are exclusively talking about network ‘proxies’ rather than any sort of responsible data movers. The networking layer is completely abstracted from the CommServe - so it can’t tell which network route is being used to achieve the backup or restore. Each individual client/MA figures that out based on the configuration so the CS is not involved in the brokering at all. So I do not think from the CS/job controller view you could know which network gateway is in use for a particular operation.
  1. This is a lot more complicated and their could be several reasons why its occurring, and then again depends on the type of workload you are restoring. In theory it should only pick an online proxy that is available - but my guess is that perhaps your proxies may also me media agents that hold index for the job. We want to avoid moving and rebuilding the index on another node unless its absolutely necessary. I could be off-base on this but it would need a much closer look at the configuration and workload to figure out what is going on.

 

J
Rank

Yes, I agree with Damian. In most cases there is a confusion between network proxies, VSA proxies (virtual server agent access nodes) etc.
The network proxies from our perspective are Commvault clients installed with FileSystem Core agent and dedicated as network proxies.

In your case, having network proxy enabled VSA proxy and media agents installed on all nodes, it may be confusing what failed. It was not the network proxy but Media Agent originally assigned for a job.

A
Rank
Userlevel 1
Badge +6

Hi guys,

Thanks for you response's. The proxies are Commvault clients installed with FileSystem Core agent and dedicated as network proxies. The MA’s are the HSX Clusters.

 

For VSA backups I originally had the HSX Nodes as my VSA Proxies in my Virtual Client. As I wanted all traffic going through the network proxies I then added the VSA client package to the network proxies and the replaced the HSX nodes with the network proxies in the Virtual client like below.

 

Reply


Terms & ConditionsCookie settings