HPE EVA Storage
1753834 Members
8407 Online
108806 Solutions
New Discussion юеВ

EVA 4400 Vdisk presentation to multiple servers

 
SOLVED
Go to solution

EVA 4400 Vdisk presentation to multiple servers

Recently, I presented a VDISK to two different servers running Windows Server 2008. After a few days I noticed that some files did not appear in one of the servers and that the filesystem seemed to be inconsistent.
I would like to know if is possible to share the same VDISK between two or more servers? If so, what is the right way to do this ?

Thanks in advance
4 REPLIES 4
Torsten.
Acclaimed Contributor
Solution

Re: EVA 4400 Vdisk presentation to multiple servers

An OS must be cluster aware if you present a disk to multiple servers at the same time.
Otherwise you see what you see now.

Hope this helps!
Regards
Torsten.

__________________________________________________
There are only 10 types of people in the world -
those who understand binary, and those who don't.

__________________________________________________
No support by private messages. Please ask the forum!

If you feel this was helpful please click the KUDOS! thumb below!   
Stu Gepp
Occasional Advisor

Re: EVA 4400 Vdisk presentation to multiple servers

If you have two separate systems writing to the same disk and some point, usually pretty quickly, each system will be updating the index file (or superblock or whatever it is called in Windows) at almost the same and one will overwrite what the other has just written.

This can be done safely with clustered systems that collaborate for FS locks but separate systems will cause FS corruption.
Ant├│nio Arag├гo
Occasional Contributor

Re: EVA 4400 Vdisk presentation to multiple servers

Use a cluster file system. See more in:

http://en.wikipedia.org/wiki/Cluster_file_system

Viveki
Trusted Contributor

Re: EVA 4400 Vdisk presentation to multiple servers

Yes, you can present same vdisk to multiple hosts simultaneosly.

If cluster systems are sharing a common LUN, only one of the node (active node) will have the access. This is taking care by the cluster itself by using some locking algorithms.