Prioritized storage unit selection based on client IP ranges.
Created: 10 Aug 2009 | 4 comments
It's not possible today, to control how backups flow across WAN lines having multiple datacenters in a single Netbackup domain, unless is manual controlled with multiple policies (and that's a nightmare).
I would like to see a prioritized storage unit selection based on client IP ranges in the policy configuration. That would enable the Netbackup admin to configure a "preferred" media server, for a IP range of Netbackup client, limiting the traffic over the WAN. If the media server is full, the next media server in the list selected.
idea Filed Under:
Comments
You can do that already by
You can do that already by setting up a STU with the 'failover' option.
- "Your backups are only as good as your restores."
You didn't get the picture.
No you can't !.
We are talking media server selection based on client IP address ranges to prevent backup traffik across WAN lines during normal operation. Let's say we have two data centeres with 250 clients and 2 media servers each. How will you prevent Netbackup from selecting a STU in the opposite datacenter ?? (no - seperate policy' and on-demand STU's mixture is not the path to go, is manual and time consuming and set limitation on recources).
If it was possible to prioritized wich client picked what media server based on IP ranges you could do somthing like this:
Range 10.1.1.1. - 10.2.2.2 use MEDIASERVER1_EAST, MEDIASERVER2_EAST, MEDIASERVER3_WEST
Datacentert EAST has IP range 10.1.1.1 to 10.2.2.2 - When Netbackup then schedule a backup of client with IP adress 10.1.2.1 it would select MEDIASERVER1_EAST or MEDIASERVER2_EAST because those media server are in the same range. If both are busy/down then MEDIASERVER3_WEST is selected, and traffic would flow across WAN lines (and that would be O.K since we wan't to secure data).
It like storage unit group - just with a evulation of client IP's.
Assumption is the mother of all mess ups.
If this post solved you’re questions please send a gratitude by marking it as a solution.
Sorry, jumped ahead in the
Sorry, jumped ahead in the post I guess.
So, how about just creating two STU groups, one per DC, and just point the local clients to the STUG with the correct media server order (local1, local2, remote1, remote2) ? Setting the STUG option to 'failover' again, of course. =)
Doing IP-specific seems like it would involve an awful lot of extra work and would have to be manually kept up-to-date to make sure it stayed useful (i.e. what happens when the network guys decide to add another backup subnet without telling you - by the time you've updated all your routing statements on a rush basis, are you going to remember to update your STUG too ? ).
- "Your backups are only as good as your restores."
Fine suggestion, but it
Fine suggestion, but it only works if you have one client per policy,it doesn't work if you have 900 client spread out in multiple datacenters in one policy ( and yes I got multiple policies with more than 100 client in them).
Being IP specific isn't the perfect solution,it will only limit the traffic going over WAN, I know that , but it is much better than today where we have no control at all. I also don't think the task of keeping the LAN table updated is a big job - target is limiting traffic, not being 100% accurate at all times.
Assumption is the mother of all mess ups.
If this post solved you’re questions please send a gratitude by marking it as a solution.
Would you like to reply?
Login or Register to post your comment.