Unsolved
This post is more than 5 years old
1 Rookie
•
32 Posts
0
1256
January 1st, 2017 01:00
UseCase for ViPR CTR v.s VPLEX Local
I'm trying to create a situation where i have
2 systems of Unity AFA 600 at the same site connected to VMware envrioment
where between them there is same copy of data.
I want to create a VMFS for example of 2TB but behind the scene it should be 2 Luns of 2TB paired in case of one of the Unity system fails.
What should be used , ViPR Volumes with Raid10 usage or VPLEX Local ? NOTE: Systems are ISCSI not FC
No Events found!
Tsky
1 Rookie
•
32 Posts
0
January 2nd, 2017 04:00
nobody knows any answer
julien3192
38 Posts
0
January 2nd, 2017 05:00
I see one main issue in your case: VPLEX in iSCSI is not supported by VIPR-C.
You could have used VPLEX mirrors (="continuous copy" in VIPR-C) if the VPLEX was in FC mode.
Newday3000
57 Posts
1
January 2nd, 2017 09:00
VIPR-C is not an array, you don't mount LUNS on VIPRC, it creates luns on real arrays. The features of the luns it creates are dependant on what the array can do.
VIPRC can create local mirror on VPLEX and host mounts VPLEX front end ports (not VIPRC is not in the data path)
You need VPLEX for this use case
Or simpler and cheaper use VMware SDRS with two datastores one on each array, setup auto vmdk placement to load share under working conditions. Half your VM's are booted from one lun (array A), the other half on the other lun (array B), replicate the lun on schedule from A to B and B to A, using mirrorview replication on regular basis, manual steps to represent replicated lun to hosts and import vm's to boot them from the failed array.
Tsky
1 Rookie
•
32 Posts
0
January 2nd, 2017 09:00
But Why VPLEX is needed? Isn't here a way to create via ViPR controller a Volume that is created via 2 different Luns behind the scene that are mirrored ?
Newday3000
57 Posts
0
January 2nd, 2017 13:00
everyone wants the same thing solved but...
array hw solution = vplex
software = vmware metro cluster (vplex or other arrays supported). The failover scenario's that must be handled are not trivial issues.
Can also build datastore on Scale IO and have more HA options but it acts as single array not two different arrays.
https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2007545
Tsky
1 Rookie
•
32 Posts
0
January 2nd, 2017 13:00
I understand what you say, but im my scenario it will not help. I want to eliminate the possibility that failed array will stuck the esxi servers. So there must be a different solution for this.