VMware Cloud Community
baber
Expert
Expert
Jump to solution

Object distribution and host failure

I have created a VM with FTT=0 according to attach pic :

vmdk component has resided on lhost4

home name space component has resided on lhost3
swap component has resided on lhost4

 also my vm has resided on lhost7

1- want to know. Which host is the main that if fail my vm will fail (except lhost4)? because at the first step I rebooted lhost3 and not happened

2- For second scenario imagine  hard disk component has resided on lhost4

vm home component has resided on lhost3

swap component has resided on lhost5

and vm has resided on lhost7

Which host is the main that if fail my vm will fail (except lhost4)?

 

Please mark helpful or correct if my answer resolved your issue.
1 Solution

Accepted Solutions
TheBobkin
Champion
Champion
Jump to solution

@baber, When you store data as FTT=0 it will place components based on what disk on what node has the lowest usage - if you want data to remain accessible after/during a disk/Disk-Group/node is offline then store the data as FTT=1, otherwise you are just making fairly pointless guess-work about what will become inaccessible when which node is offline.

 

When you rebooted node lhost3 it probably only didn't become inaccessible because nothing was actively writing to the namespace and the vmx data was already in memory - if you tried vMotion or take snapshot (or a lot of other things) of a VM in such a state it would become inaccessible (because the namespace cannot be written to as it would be inaccessible).

 

I don't really understand the point of your other points - either store data with redundancy or don't (and be prepared for it to be unavailable or necessary to restore from backup when it is inaccessible).

View solution in original post

1 Reply
TheBobkin
Champion
Champion
Jump to solution

@baber, When you store data as FTT=0 it will place components based on what disk on what node has the lowest usage - if you want data to remain accessible after/during a disk/Disk-Group/node is offline then store the data as FTT=1, otherwise you are just making fairly pointless guess-work about what will become inaccessible when which node is offline.

 

When you rebooted node lhost3 it probably only didn't become inaccessible because nothing was actively writing to the namespace and the vmx data was already in memory - if you tried vMotion or take snapshot (or a lot of other things) of a VM in such a state it would become inaccessible (because the namespace cannot be written to as it would be inaccessible).

 

I don't really understand the point of your other points - either store data with redundancy or don't (and be prepared for it to be unavailable or necessary to restore from backup when it is inaccessible).