Solved

How to Add a Node to Content Indexer

  • 8 April 2021
  • 8 replies
  • 69 views

Badge +1

That’s it.  I see that I’m at the limit of indexed items for a single node, and our indexed material continues to grow.  Email in this case, but my file indexer will be full on it’s first node within the next few months.

How do I add another node?

icon

Best answer by arunprasada 8 April 2021, 21:16

Odi,

 

You are currently on our old exchange mailbox agent and using the old way of content indexing the data using what we call Search Engine. Search engine has a limit of 30 millions emails per node. You can add new nodes to the current engine and keep continuing. But that is not going to be my recommendation.

 

We would like you to start using the new Exchange agent and the new way of indexing. 

https://documentation.commvault.com/11.23/essential/125239_getting_started_with_exchange_online.html

Here is the steps to follow do the transition.

https://documentation.commvault.com/11.23/expert/29238_transitioning_to_exchange_mailbox_agent.html

Once you move, the index server which you will be used to store the emails moves from 30 million to 250 million based on the machine configuration. Details here

https://documentation.commvault.com/11.23/expert/28819_hardware_recommendations_for_index_server.html

 

You have 30 million content indexed emails and they may have to be reindexed to get this bigger capacity. You will have plan for that.

Finally, index servers can run on VMs. Just make sure they have dedicated CPU and memory. 

 

View original

8 replies

Badge

Odi,

 

You are currently on our old exchange mailbox agent and using the old way of content indexing the data using what we call Search Engine. Search engine has a limit of 30 millions emails per node. You can add new nodes to the current engine and keep continuing. But that is not going to be my recommendation.

 

We would like you to start using the new Exchange agent and the new way of indexing. 

https://documentation.commvault.com/11.23/essential/125239_getting_started_with_exchange_online.html

Here is the steps to follow do the transition.

https://documentation.commvault.com/11.23/expert/29238_transitioning_to_exchange_mailbox_agent.html

Once you move, the index server which you will be used to store the emails moves from 30 million to 250 million based on the machine configuration. Details here

https://documentation.commvault.com/11.23/expert/28819_hardware_recommendations_for_index_server.html

 

You have 30 million content indexed emails and they may have to be reindexed to get this bigger capacity. You will have plan for that.

Finally, index servers can run on VMs. Just make sure they have dedicated CPU and memory. 

 

Badge +1

This link you’ve just shared also indicates 30 million objects shouldn’t be a factor here.

If I recall correctly, we would’ve specified this server according to the “Medium” sizing contained in that Hardware Requirements link, which shows as 150 million items *per node*.

Do not see any limitation as far as running it on a virtual vs physical. 

Make sure to follow the specs above and odd number of node requirements.

Note: The HAC cluster must contain an odd number of nodes. For small to medium environments, use three nodes. For large environments, use five nodes.

Hardware Requirement: https://documentation.commvault.com/commvault/v11/article?p=28819.htm

 

Badge +1

Am I misinterpreting this situation?  It doesn’t look like 30 million objects alone is the problem here.

We don’t need HA.  Total number of emails is currently barely over 30 million.  I don’t even know what Command Center is; fairly certain we’re not using it.

==============

Use a multinode index server environment only in the following conditions:

  • If high availability of index server is a requirement.
  • If number of emails (Non Content Indexed objects) is greater than 500 million.
  • If number of emails (Content Indexed objects) is greater than 250 million.
  • If you are using target index server for Case Manager in Command Center.
Badge +1

I see now that CommVault is simply misusing the word Cloud to describe an HA configuration for a local datacenter, or so it seems.

Is there a recommendation to not run these nodes on virtual architecture?  i.e. Do I need to spec out a group of physical CI servers?

Badge +1

We have two separate indexing engines at the moment, hosted completely separately on physical machines(dedicated CI media agents).  The one that is full on it’s first node at the moment is only for email.  Bearing in mind, when I say “full”, the only limitation I currently see is the 30 million items limit being enforced by the CommVault software.  The hardware still has plenty of space, and we have not noted any decreases to performance, other than that email indexing is currently not running at all.

We started these indexing engines in 2016.  We currently have no plans to stop indexing all email, with the possible exception of non-business related emails if there’s a way to filter those out of the indexer automatically(I have never tried, but regardless this would only prolong the problem slightly).

Unfortunately, cloud services is completely out of the question.  Our WAN bandwidth is a meager 20 megabits per second, and it barely supports our day to day activities, without adding infrastructure level data transfer.

If there is not a way to implement additional CI nodes on physical architecture within our datacenter, we will likely be forced to find a new CI software provider.

Badge

Hi Odi,

Can I get more details on your setup before I suggest a solution? 

What data are you indexing? Email is guess but u also refer to files

How many documents that the current index server have?

When did u start indexing ?

You are likely looking to add another node and to achieve that you would be looking at the Index Server Cloud. 

https://documentation.commvault.com/commvault/v11/article?p=97994.htm

Multinode Index Servers Using Index Server Cloud

Use a multinode index server environment only in the following conditions:

  • If high availability of index server is a requirement.
  • If number of emails (Non Content Indexed objects) is greater than 500 million.
  • If number of emails (Content Indexed objects) is greater than 250 million.
  • If you are using target index server for Case Manager in Command Center.

The steps to achieve that are laid out in the following link: https://documentation.commvault.com/commvault/v11/article?p=97931.htm

 

So, aside from deploying a new Index Node, I would strongly recommend that you revisit the environment because the index will only grow as users receive more emails. For instance, do you need to content index all users? or could you set CI for a group of users thus reducing the amount of data being content indexed. 

Hope that helps.

 

Reply