Replicatation and Protect stgpool

[StiffBoard]

If I want to replicate my VM backups, Spectrum protect for VE and server is 8.1. Do I need to Protect stgpool and then replicate *? All backup goes to directory pool container on each server.

With only Protect stgpool the nodes will not be defined on the other server? How do I go about that? Do I need to create all the VM backup nodes? Data mover, Proxy, datacenter nodes and so on?

 

[marclant]

Before you enable protect and replication, you have to make sure your configuration can handle it. Make sure that you are using SSD for the database and activelog on both the source and target. Determine your daily ingest (amount of TB backed up daily) and find out which category you fall into with the blueprint (small/medium/large) and make sure your system is closely matched to the blueprint. The hardware doesn't have to be the same, but it has to be a close match with similar specs. https://ibm.biz/BdHFmv

Do I need to Protect stgpool and then replicate *?

PROTECT will make protect your storage pool. So it will copy all the deduplicated extents from the source server and copy them to the target server. If your storage pool gets damaged, you can recover it from the data on the target.

REPLICATE NODE does 2 things. If copies the extents if protect didn't already copy them, and also replicates the metadata so that your clients can restore data from the target should fail.

Technically, you could get by with just replicate node, however protect stgpool is a lot more efficient at moving data than replication. So always run protect first, followed by replicate. NEVER RUN THEM AT THE SAME TIME, it will bring your DB to a crawl with all the DB updates for the same data.

With only Protect stgpool the nodes will not be defined on the other server?

Correct

How do I go about that?

Replicate

Do I need to create all the VM backup nodes? Data mover, Proxy, datacenter nodes and so on?

On the source yes, not on the target, replication will do that for you.

 

[StiffBoard]

Well, I'm assuming right then, but it's not happening for me. Just discovered q replnode, but guessing I'm not supposed to define a replnode? That is done by the system I guess.

Not very big Environment at the moment so I ran a protect storage, and now I can see Capacity used is about the same on both servers, that seems right. Started replicating a machine that was previously not replicated, and it starts copying about 400 GB?! Shouldn't data already be in Place? q occ on that machine is around 450 GB

But for my Spectrum for ve nodes, which I have set to replicate even though filespace for VMs only reside on node for vm datacenter, not much is happening. Get these failures which I assumed would be automatically fixed by defining the nodes.

15:03:00 ANR1929E Replication of node VMBKP_VMDATACENTER to target replication server SPECTRUM2 failed. The replication mode of the source node is SYNCSEND, and the node is not defined on the target replication server.

And process ends with failure.

 

[marclant]

It can get real complicated really quick.

15:03:00 ANR1929E Replication of node VMBKP_VMDATACENTER to target replication server SPECTRUM2 failed. The replication mode of the source node is SYNCSEND, and the node is not defined on the target replication server.

Issue: HELP ANR1929E and follow the recommendations.

Started replicating a machine that was previously not replicated, and it starts copying about 400 GB?!

I don't recall if it's referencing the logical bytes (the size of each file) or the actual amount of data transferred. Two different things.

 

[marclant]

Not very big Environment at the moment

They all start out that way and then creep up beyond what it can handle.