[SOLVED] Expand Cluster or keep seperate

[Haider Jarral]

Hello Experts,

Pardon my ignorance.
I am in process of migrating my existing data centre physically to a different location and at the same time adding 3 more nodes. I have done a lot research but can't reach a conclusive answer.

This is what I have right now

4 X SuperMicro Server, 128G RAM Each and 8TB HDD (This is all in Ceph Cluster with 2x1G cluster network)
4 X HP Proliant 360 G7, 72G RAM Each with 1TB SSD (This is also in a Separate Ceph Cluster with 2X1G network)

My question is if I want to combine these two setups into a single Ceph Cluster, how would I setup the storage.
I want to have a pool available for fast VMs using only SSDs and I want to have a pool available for slow VMs using only HDDs. Also Can I host VMs with fast pool on server with HDDs or vice versa ? What would happen in case of failover when VM with fast pool migrate over ?

If I combine these two setups in cluster, how easy it is to accomplish above mentioned scenario.

Or Am I just better of keeping these cluster separate.

 

[Haider Jarral]

Bump bump, weekend is drawing near and I have to decide to either tear down or keep intact. Any input is appreciated. Thanks

 

[sg90]

I guess the best question is, do you need the data that is currently in these two CEPH environment's or can you nuke the environment fully and start fresh?

 

[Haider Jarral]

I am already taking full backups on NAS so yes I can nuke it, If nuking is the way to go.

 

[sg90]

Then yes you could nuke and add all the nodes/OSDs to one CEPH Cluster and just run two separate pool's one for SSD and one for HDD.
Would mean you only need one set of mons and one dashboard/monitoring view.

 

[Haider Jarral]

This sounds like its doable, can you please provide a reference link or video. Thank you for all your help.

 

[Haider Jarral]

Should I mix HDD and SSD in each server or the crush rules I create will allow me to create a VM on Super Micro and still use SSD pool ?

 

[spirit]

you can create 2 crush rules and assign them to 2 different pool


ceph osd crush rule create-replicated replicated_ssd default host ssd
ceph osd crush rule create-replicated replicated_hdd default host hdd

ceph osd pool set pool1 crush_rule replicated_ssd
ceph osd pool set pool2 crush_rule replicated_hdd



(ceph detect ssd/hdd, so you can mix them where you want)

 

[Haider Jarral]

What about this case then

Can I host VMs with fast pool on server with HDDs or vice versa ? What would happen in case of failover when VM with fast pool migrate over ?

 

[Haider Jarral]

I got it all figured out. Thanks a lot for your help and tremendous guidance from @Seed.