Solved

Native replication (RF2, RF3) vs Local Replication (Async DR, NearSync)

  • 7 December 2019
  • 4 replies
  • 10922 views

Badge

Hi all,

 

Local Replication is a process in which multiple copies of data are stored within a storage container. These copies exist for fault tolerance. Snapshots are placed locally on the same cluster as the source VM. Thus, If a physical disk fails, the cluster can recover data from another copy. The cluster manages the replicated data, and the copies are not visible to the user.

 

So, what is the difference the Replication Factor option? Because RF is used too for fault tolerance in case of a physical disk failure (or node, ...)

 

Thanks

icon

Best answer by HITESH0801 7 December 2019, 12:45

Hello @Sam41 
Local Snapshot is used if you want to restore your VM to a previous version using local snapshot, but RF copies are used for fault tolerance and are used to rebuild the data when a node or disk goes bad.
Replication of data happens at every write whereas snapshot get’s taken at a particular schedule.

 

Let’s consider two scenarios

1 → You have a local snapshot saved, and your VM got corrupted and you want to go to a previous version, you restore your VM using local snapshot. 

2→ One of your physical disk goes bad, the data which was in the disk get’s rebuild(all happens behind the scenes )   and you have no data loss and your VM is totally fine. This happened because a duplicate copy of the data in the disk was stored somewhere else in the cluster and was used to rebuild the data. (Replication Factor in play). 

View original

4 replies

Userlevel 4
Badge +5

Hello @Sam41 

Help me understand your query.
According to the latest Prism Guide.
 

 

local replication

Multiple copies of data within a storage container. These copies exist for fault tolerance: if a physical disk fails, the cluster can recover data from another copy. The cluster manages the replicated data, and the copies are not visible to the user.

 

https://portal.nutanix.com/#/page/docs/details?targetId=Web-Console-Guide-Prism-v511:Web-Console-Guide-Prism-v511

So basically, RF is saving multiple copies of data in your cluster for fault tolerance, and async DR and Near sync involves saving a copy of your data via snapshot to another cluster/remote site for disaster recovery and backup. 

This post might help you understand RF-2 and RF-3 better.
https://next.nutanix.com/how-it-works-22/confused-regarding-rf-2-and-rf-3-and-which-one-is-suitable-for-your-cluster-33415

Regarding Backup and Disaster Recovery, try giving the section “Backup and Disaster Recovery” a read from out Nutanix Bible 
 

 

 

Badge

 

local replication

Multiple copies of data within a storage container. These copies exist for fault tolerance: if a physical disk fails, the cluster can recover data from another copy. The cluster manages the replicated data, and the copies are not visible to the user.


So basically, RF is saving multiple copies of data in your cluster for fault tolerance, and async DR and Near sync involves saving a copy of your data via snapshot to another cluster/remote site for disaster recovery and backup. 

 

Thanks for your reply.

Async DR and Near Sync both save a copy of data via snapshot locally (Time Stream) and remotely.

My question was: Why local snapshots when we already have Replication Factor?

Userlevel 4
Badge +5

Hello @Sam41 
Local Snapshot is used if you want to restore your VM to a previous version using local snapshot, but RF copies are used for fault tolerance and are used to rebuild the data when a node or disk goes bad.
Replication of data happens at every write whereas snapshot get’s taken at a particular schedule.

 

Let’s consider two scenarios

1 → You have a local snapshot saved, and your VM got corrupted and you want to go to a previous version, you restore your VM using local snapshot. 

2→ One of your physical disk goes bad, the data which was in the disk get’s rebuild(all happens behind the scenes )   and you have no data loss and your VM is totally fine. This happened because a duplicate copy of the data in the disk was stored somewhere else in the cluster and was used to rebuild the data. (Replication Factor in play). 

Badge

Hi @HITESH0801 ,

thanks for your clarification with the use cases scenarios.

 

In addition to your reply, I read that the data in the snapshots is not copied or moved.

I was thinking that the snapshot will create another copy of the data.

https://nuschool.portal.relayware.com/content/CourseWare/2055/scormcontent/index.html#/lessons/FMpwA6UYQ4DVDQhtwU51z8ZuFjm8zU2g

 

Thanks.

Reply