VMware Cloud Community
psix
Contributor
Contributor

Cluster storage for vmware vSphere.

HI,

i have 3 servers with 1 TB space on each,

How can i build redundand storage for my virtual infrastructure? I need redundancy like if one server go down other server are working.

And, if this possible, need balance Active Active Active Cluster...

I think about OpenSolaris but i dont undestand how build without using other servers for concentrate ISCSI targets...

Can you help me plz ?

Reply
0 Kudos
12 Replies
kac2
Expert
Expert

Local storage is not going to give you the availability to want. When it comes to the virtualized world, local disks are bad.

You need to invest in a SAN. There are an unlimited amount of SAN vendors. time to start researching.

Reply
0 Kudos
golddiggie
Champion
Champion

You need a NAS/SAN for the VM's to reside upon in order to get there. Look at products like the EqualLogic PS6000 and PS6500 series, or even the PS4000 series if you have lower IOPs needs. There are others, like the Dell MD3000i too.

You'll want to get the array with enough storage to at least provide you with the same amount you have now (3TB usable). I know the 4TB solution from EqualLogic actually gives you about 2.5TB of usable space.

Whatever you end up with, I would go with an iSCSI solution (the ones above are iSCSI).

With those, the VM's actually reside upon the LUNs that the ESX/ESXi hosts connect to. Then, using HA and/or DRS you can set it so that if any host fails, the VM's are picked up (quickly) by either of the remaining online hosts. Depending on the host servers, and the VM's, you might even be able to get away with everything running on a single host, in case two fail. I wouldn't recommend it, and only do that in a dire emergency.

With HA and DRS, ESX/ESXi will load balance the VM's across all your hosts. It might look funny to you, at first, but you'll soon see how it's logic actually makes sense.

VMware VCP4

Consider awarding points for "helpful" and/or "correct" answers.

Reply
0 Kudos
psix
Contributor
Contributor

I dont need hardware solution, i need configure this server to make cluster NFS or somthing like this.

Reply
0 Kudos
kac2
Expert
Expert

Psix, I don't think you understand how HA works. For your servers to be able to fail over to one another, the VMs must reside on centralized storage that is accessible to each anmd every server.

If you have a VM1 sitting on ServerA, and ServerA has a failure, ServerB cannot access VM1 because it resides on ServerA's local storage.

Once the VM1 is centrally located on a SAN or NAS, it is accessible by both ServerA and ServerB, therefore making failover possible. Does this help you understand?

Centralized storage can use protocols such as iSCSI, NFS, Fiber Channel, FCoE, etc. you need take some time and research SAN vendors. I personally would go with NetApp, Compellent, EMC, Xiotech, just to name a handful.

jkumhar75
Hot Shot
Hot Shot

The feature which you are looking it out for the redundancy purpose in VMware ESX cluster it's possible only with the shared storage.Without shared storage you can't make use of the HA & DRS feature which you are looking it out.

Jay

MCSE,VCP 310,VCP 410

Consider awarding points for "helpful" and/or "correct" answers.

If you found this or other information useful, please consider awarding points for "Correct" or "Helpful". Jayprakash VCP3,VCP4,MCSE 2003 http://kb.vmware.com/
Reply
0 Kudos
kac2
Expert
Expert

Psix,

are you trying to cluster these 3 servers together to be your centralized storage and the VMs are sitting on a different set of servers? If so, you might want to check out a product called Datacore SANmelody. but i'm not even 100% sure if that will solve your problem.

Please award points for correct or helpful answers.

Reply
0 Kudos
psix
Contributor
Contributor

You not undestand me, i need cluster shared stoprage for esx. I dont need clustered ESX, esx are working on other hosts.

I need somthing like lustre, GFS or other..

Reply
0 Kudos
VMFrankL
Contributor
Contributor

Hi , I believe that you are looking for something like replication data over nodes.

I am using the glusterfs , is very easy to use and maintain using ubuntu linux servers.

http://www.howtoforge.com/striping-across-four-storage-nodes-with-glusterfs-on-ubuntu-9.10

http://www.gluster.org/

please feel free to ask me any question.

Regards

Frank

Reply
0 Kudos
Silverchenau
Enthusiast
Enthusiast

Your question is more like answered by hardware solution of DR site and SAN replication.

It's very hard to just use software to keep sync.

I personally recommend EMC Recover Point, but again, it's hardware&software solution.






- Silver

My Vmware blog: http://geeksilverblog.com

- Silver My Vmware blog: http://geeksilverblog.com
Reply
0 Kudos
Josh26
Virtuoso
Virtuoso

You not undestand me, i need cluster shared stoprage for esx. I dont need clustered ESX, esx are working on other hosts.

I need somthing like lustre, GFS or other..

Do you mean you want to install two VMs and run GFS on both of them? You don't need us for that.

If I read you correctly and you are wanting ESX to mirror its datastore, then you are asking about something ESX does not do - you need a SAN.

Reply
0 Kudos
AntonVZhbankov
Immortal
Immortal

You can use HP Lefthand or Starwind iSCSI to build redundant virtual SAN on top of local storage.


---

MCSA, MCTS Hyper-V, VCP 3/4, VMware vExpert

http://blog.vadmin.ru

EMCCAe, HPE ASE, MCITP: SA+VA, VCP 3/4/5, VMware vExpert XO (14 stars)
VMUG Russia Leader
http://t.me/beerpanda
Reply
0 Kudos
Gluster
Contributor
Contributor

Frank,

Thanks for your comments.

We'd love to get your feedback on the new  VMWare virtual appliance version of the open-source GlusterFS platform for scale-out storage.

The product is intended to enable easier deployment and use of Scale-Out NAS in a VMWare environment.

The product is availabe for use as a 30 day free trial. We are looking for feedback on features, performance, and overall utility.

The Gluster Virtual Storage Appliance for VMware environments can be deployed on any VMware certified host server hardware platform and certified storage: JBOD, DAS, or SAN. High availability requirements in VMware environments are fulfilled via n-way replication capabilities on all Virtual Appliances.

Gluster is POSIX compliant, runs as a virtual machine, and does not require applications to be rewritten.  Users of VMware and the Gluster Virtual Storage Appliance can aggregate CPU, memory, and capacity in one Global namespace, and scale-out capacity and performance on-demand. The system is designed to provide better storage economics via the use of commodity hardware. Clients can access via the use of NFS or via Gluster's native FUSE client.

For more information, or to download, please go here: http://bit.ly/eiE50N

Detailed performance information can be found here: http://bit.ly/hI0aDh

Again, any feedback--positive or negative--would be appreciated.

Reply
0 Kudos