Unsolved
This post is more than 5 years old
32 Posts
0
1383
New Install of Networker and 2 DataDomain 670's
I'm looking for some advice on implementing a new roll out. I want to build a new Networker/DataDomain infrastructure from scratch. I'm trying to figure out at a high level scenario of how to implement it. Below is my current setup. I could upgrade but i'm looking for a fresh start. I backup roughly 200 VM's. Friday fulls/daily incrementals.
Main site - (1) virtual 2008r2 Networker server with all componets installed. (1) DD670 DDBOOST configured.
DR site - (1) DD670 DDBOOST configured with clone replication to main site. (no networker servers in this location).
My main goal is to make my DR site my production Networker NMC server. I would like both DD670's to replicate to each other and have the ability to recover backups if any location may go down. Would this mean building the following .....
DR Site - (1) NMC (1) Storage node (1) Server . Should i split these up into seperate servers like NMC and Server/SN perhaps?
Main Site - (1) Storage node (1) Server all one server or split it up?
Any guidance would be great because i've always felt my initial Networker setup i've been running for years now wasnt done correctly.
IMPORTANT: I have all the ESX virtual resources and SAN storage to build this out correctly. I don't need combine features together if un-needed.
ble1
14.3K Posts
0
June 25th, 2014 13:00
Do you have vmotion between main and DR site?
actadmin1
32 Posts
0
June 25th, 2014 14:00
No, the ESX data center at the DR site is stand alone from our main site.
actadmin1
32 Posts
0
June 26th, 2014 08:00
In simple terms. I have 2 sites and 2 DD670's. 100Mb reliable ethernet link between sites. What's the best way to setup the servers at each location?
crazyrov
4 Operator
4 Operator
•
1.3K Posts
1
June 27th, 2014 01:00
Here's my suggestion :
Production Site :
1 - NetWorker Sever
1 - Storage Node (Also make it as a NMC server)
DR Site :
1 - NetWorker Sever
1 - Storage Node (Also make it as a NMC server)
Enable clone based replication on DD and on NetWorker create clone devices from the DR site DD as target. This way NetWorker will have all the media information required. Enable bootstrap notification to ensure that you have al the latest bootstrap information for you site.
The VM's backups are done using VADP or as regular client level ?
Thierry101
326 Posts
1
June 29th, 2014 16:00
if you gn to make DR Server as active then that will be different name from Prod server name.
If Prod goes down or DR goes down then you cant recover the clients since clients will look for it''s backup server name which is unique. You probably have to do active active Networker cluster...?
actadmin1
32 Posts
0
June 30th, 2014 08:00
Thanks for the information so far everyone.
@Crazyrov I like your setup idea. Why would i need to 2 NMC's? Arent you supposed to have only one per datazone? and don't you put the NMC with Server and leave the storage node alone? I see you suggested putting the NMC and Storage node together.
actadmin1
32 Posts
0
June 30th, 2014 09:00
Here's the setup I'm thinking. Lets forget about DR/Production and just think Site1 and Site2
Site 1 - NMC/Server and Storage Node - DD670 w/clone replication to site2
Site 2 - Server and Storage Node - DD670 w/clone replication to site1
If site 2 is wiped out completely i would still have the NMC and all the replicated backups at site1.
Can i actually do anything with this replicated data though? Could i rebuild site 2, get a new Server/Storage/DD670 and rebuild from the replicated data that was safely located at site1?
ble1
14.3K Posts
0
June 30th, 2014 17:00
I like approach with active/active or active/passive site approach. You seems to have sites isolated and wis to replicate data between sites (old fashioned "DR site" approach). My first question to DR site approach is the one you already asked - if site X goes down, do you have necessary hardware to rebuild it? If not, you only have archive and not backup really. NMC is trivial (you can have multiple per datazone and/or one for multiple datazones - installing one takes minutes so it is not so important unless talking about a bit more complex setups where your infra uses some VM integration or hosts multiple VLAN which are needed for DD setups and devices). Anyway, with setup you listed, you can't do much. Why? Because serves on site A and B are independent of each other (the way you wrote it above) and they just replicate to other site. You would need to scan in records from failed site first to be able to start to use data.
crazyrov
4 Operator
4 Operator
•
1.3K Posts
0
July 1st, 2014 01:00
I leave most of the load off the NetWorker server including the NMC server, its a practice that i follow, no much gain from it though. 2 NMC's because you need an NMC to access you DR site when your production is destroyed.
ble1
14.3K Posts
0
July 1st, 2014 02:00
It would be fair to say that EMC best practice guide itself says that if you have more than 50 clients I think you should also skip it in your case as you already have 200 VMs.
actadmin1
32 Posts
0
July 2nd, 2014 07:00
Thanks everyone for your great help. I'm working on a plan now withe the information i've gathered.