Unsolved
This post is more than 5 years old
1 Rookie
•
22 Posts
0
945
February 28th, 2017 14:00
"Scaled-up" 3-VM limitations
The support matrix mentions a scaled-up deployment, but it is not real clear on what is meant by “scaled up.”
Also, it states a 3-VM deployment can handle 20% of the maximum. So, if a 5-VM can handle 170 block storage systems, a 3-VM can only handle 34? That doesn’t sound right.
No Events found!
isakats
141 Posts
0
March 1st, 2017 05:00
Hi bellitch,
"scaled up" means that you can increase the size of the SRM environment in order to monitor more devices; 34 block on a 3 VM setup sounds about right as it would have a single database.
Note that this can vary, a single database can store 1.5 million metric, the number of metrics is based on the number of components such as luns, ports, datastores, vm's, etc...
hth
regards,
Isaka
julien3192
38 Posts
1
March 1st, 2017 06:00
I think the question is for VIPR-Controller (not SRM).
"20% of maximum" is the scalability limit, that means a VIPR-C 3-VMs can go to 80% of the 5-VM limits, that is 136 Block arrays for instance.
If you plan to use VIPR in a Production environment (especially with >30 storage arrays !), we strongly recommend to deploy a 5-VM instance (better protection, redundancy...etc).
julien3192
38 Posts
0
March 1st, 2017 07:00
To go from 3->5VM (or 5->3 VMs), you can restore a VIPR backup using the procedure detailed here (page 41+):
http://www.emc.com/collateral/TechnicalDocument/docu79624.pdf
bellitech
1 Rookie
•
22 Posts
0
March 1st, 2017 07:00
Correct. I was talking about Controller. 80% of the maximum makes more sense.
As for the "scaled up," does that mean you can go from a 3 VM deployment to a 5? That part still doesn't make sense to me. I always recommend 5 VM for production, but one customer is stuck on starting with 3, for some reason.