Hello Community,
Quite new regarding SRM and I may have not thoroughly read the manual, so expect some novice-level question.
Just wanted a confirmation regarding implementation of SRM version 5.0 (may need an upgrade due to compatibility issues with storage arrays) for various vSphere clusters:
- 5 clusters vSphere 5.0 managed by a vCenter server v5.0,
- 6 clusters vSphere 5.5 managed by a vCenter server v5.5.
In case I want to implement a single set of clusters in my DR site (in version 5.5, at least), I'll have to install two SRM plugins:
- one to enable pairing with the one in vCenter 5.0
- and another one to enable pairing with the plugin in vCenter 5.5, ¿right?
Note: migration+upgrade of v5.0 VMs and upgrade of v5.0 ESXi host is already running, but very slowly due to organizational complexities, so upgrade as pre-requisite is unfortunately not an option.
Another question: have you ever seen some kind of Recovery Time estimator for SRM. Of course, I guess it would need a lot of parameters or assumptions: hw platform, number and versions of ESXi, vCenters, SRM, # and configuration of VMs, # datastores, # VMs/datastores, # VM start-up/host, level of customization with specific scripts in recovery plans, number of recovery plans, controlled or real disaster recovery plans, etc.).
Nothing is better than real-life testing, but such estimator would help a lot in some internal decision making processes.
In our situation, we would start with simple recovery plans (less than 3), handling 1.200 VMs approximately in total, with only IP@ change, VMware products versions mentioned above, platform details:
clusters v5.0: 32 Dell PE M610, 11 M620 and 7 M910, total of 114 CPUs, 336 cores, total 10,7 TB RAM, 600 VMs, 195 datastores (sigh) full FC SAN, 138 TB used.
clusters v5.5: 4 HP BL460c G8, 44 IBM/Lenovo Flex x240, total 96 CPUs, 400 cores, 18,5 TB RAM, 600 VMs, 90 datastores full FC SAN, 161 TB used.
All replication is array-based, mix of EMC VNX, HP 3PAR and IBM Storewize FC SAN arrays.
Later, we'll implement application/service-based recovery plans.
Thanks a lot for your support.
Regards,
Cyril