Solved

RF2 case with 3 node cluster and 3 disk failure in same node

  • 28 January 2021
  • 7 replies
  • 1355 views

Badge

Dear All,

May I have a question about the below situation?

RF=2

3 node cluster

2 x disks failure in same node

May i know whether the cluster will be corrupted or not? From my understanding, it should not be corrupted as 2 x disks failure in same node (NOT different node).

icon

Best answer by AnishWalia20 29 January 2021, 09:36

View original

This topic has been closed for comments

7 replies

Userlevel 6
Badge +5

Hey, @KC93 that is absolutely correct. 

The reason is that the data from those 2 failed disks will also be present on other nodes in the cluster.

So say fs a node goes down and all its disk fails too, you won’t be having data loss in that scenario on an RF2 cluster too because random other nodes will have the data copied.

I hope I cleared your doubts. Let me know If I can help you with anything else. :smile:

 

Badge

Hi AnishWalia,

 

Thanks for your information and help.

 

So, if 2 x failed disks in same node (Under RF2 situation), may i know whether the node will be down or not? or still operate? Thanks

 

Userlevel 6
Badge +5

hey @KC93 it’s a pleasure to help.

 

So it depends on whether SSDs are failing or HDDs are failing.

Say 2 SSDs on the Node go bad at the same time, then definitely the node will be down, but not in case of HDD, say 2 HDDs fail on a Node, it can still function but with less space and less chance of Node failure.

 

Badge

Hi AnishWalia,

 

Thanks for your information. Very Clear….

That’s mean if 2 x SSD down, CVM will down. May i know the default CVM location should be under SSD slot 0 right? If SSD drive in slot 0 failure, CVM will be rebooted into SSD slot 1 right?

 

And if another situation is if only CVM down, the underlying storage will become inaccessible and thus those VMs need to go to another node to get data via 10GE network, right?

 

Thanks again

Userlevel 6
Badge +5

Yes @KC93 , that is correct. :smiley:

 

Say if for some reason a CVM is down and its storage is inaccessible, in that scenario the VM’s storage/data IOs will be taken care of by other CVMs or other nodes in the cluster.

 

You can get more information about this here https://nutanixbible.com/ . Just search for Node failure.

 

 

 

 

 

 

You can also find some amazing videos on Nutanix’s You tube channel https://www.youtube.com/c/NutanixUniversity

I hope these help. 

 

Badge

Hi AnishWalia20 ,

 

Thanks for your information.

 

Userlevel 6
Badge +5

Hey @KC93 , it is always a pleasure to help. Glad you found it useful

 

Let me know if I can help you in any other way or answer any of your questions:smile: