Networker DDBoost Client Direct with Multiple VLANs
Networker server setup with different IPs for VLAN15 and VLAN16 to backup different VLANs servers
Switch level done VLAN tagging
DD setup vth0 10.x.x.x (using 2 physical interface ethMa, ethMb), then create 2 virtual vth0.15 and vth0.16 interface.
At Networker connect to DD, create Data Domain using IP address of vth0.15..
Backup to DD has no problem, but I wonder the data is still flow through Networker server or ddboost Client Direct does not work due to setup? I login to DD and show only 1 access at the time from networker server..
How is networker client redirect the traffic to DD via their VLAN respectively? Since Networker covered both VLAN15 and VLAN16, i think the client still pass through networker even i already select client direct and backup to the DD pool...
You can examine the traffic on your media server (card) and also check logs as if it can't use direct stream to DD it records so in daemon.raw. I have also different VLANs, but for those (each) I defined DD and have their own pools. In other words, I have devices for each VLAN (3 in my case) and depending on which VLAN client is, it belong to group that will use that VLAN. With such setup, I see all my clients using correct VLANs towards DD (including replication).
Yes. For example, I have server with 3 VLANs. Depending if you have 1 site or 2 sites, you will end up with 3 or 6 pools respectively (for backup; for cloning double it). Example:
- you have VLAN1, VLAN2, VLAN3
- you have only backup server and it has all 3 VLANs
- I will assume you use cloning (CCR)
- DD1 is running on site1
- DD2 is running on site2
In NMC you create DD devices for using DD1 VLAN11 (and you use DD1 VLAN name to define DD):
- DDsite1_backup_VLAN1
- DDsite1_clone_VLAN1
Then same for DD2 VLAN1 (and you use DD2 VLAN1 name to define DD).
Then same for DD1 and DD2 VLAN2 (now you use VLAN2 names to define DD)
Then same for DD1 and DD2 VLAN3 (now you use VLAN3 names to define DD).
Assign deviced to pools which will be something like:
Site1_bVLAN1
Site1_bVLAN2
Site1_bVLAN3
Site2_bVLAN1
Site2_bVLAN2
Site2_bVLAN3
Site1_cVLAN1
Site1_cVLAN2
Site1_cVLAN3
Site2_cVLAN1
Site2_cVLAN2
Site2_cVLAN3
If you have single site, you can cut this in half of course as you don't need to worry about site separation between pools. Also, if you use server plus two media servers (one per site), you can have 50% less pools as site to which it writes is then controlled by storage node field in client. Above example, one backup server + 2 sites, is something that should be very common these days, but NW still doesn't have option to instruct client to which DD to write in such case so you must separate streams per site (unless you don't care about traffic going cross site back and forth).
The setup is only 1 site with 1x DD, and 2 VLANs. No Cloning. (Only clone to tapes)
Will it seems like the DD (with 2x IPs) will be like “partition” to served 2 VLANs?
NW will have to create 2 Data Domain (with different VLANs), and each Data Domain having its own device and volumes to serve the VLANs respectively? (different MTREE inside DD)
thanks a lot for your reply, it helps a lot...It seems like a mult-tenancy environment, where using 1 DD to serve different VLANS servers, although this is same entity here.
if 2 device path and 1 device path deduplication performance will be the same? since Global Deduplication?
ble1
6 Operator
•
14354 Posts
•
56186 Points
2175
0
Posted January 21st, 2016 13:00
You can examine the traffic on your media server (card) and also check logs as if it can't use direct stream to DD it records so in daemon.raw. I have also different VLANs, but for those (each) I defined DD and have their own pools. In other words, I have devices for each VLAN (3 in my case) and depending on which VLAN client is, it belong to group that will use that VLAN. With such setup, I see all my clients using correct VLANs towards DD (including replication).