11-28-2016 06:18 AM
Hi Experts,
I have a customer which wants NBU Master Server need to be in High-Availabilty but not ready to buy cluster framework license.
So, as per consideration Backup Master is going to be deploy on VM but customer wants
VM Machine (Master Server) should reside on single ESX host (No vMotion)
No External RDM Luns to be presented to host as a Shared Disk.
How can we achieve High Availabilty in such scenerio
Regards,
Mazhar Khan
Solved! Go to Solution.
11-29-2016 12:44 AM
Please suggest what can be done in such case
Walk away.
Your customer is not serious about High Avalability.
Catalog backup is all they can rely on to protect the master server.
11-28-2016 06:36 AM
There is no such thing as High Availability without spending money.
Forget about NBU for now and consider components needed for HA:
Every single component needed to run a particular application must be redundant - server, storage, network etc, etc.
Your customer needs to make a decision BEFORE NBU master server is installed - NBU can only be clustered during brand new installation.
If they want to cluster NBU at a later stage, they will need to spend additional money for consulting services to migrate from standalone master to clustered master.
Requirements to cluster NBU is documented in NetBackup Clustered Master Server Administrator's Guide
Other HA possibilities (such as Catalog Replication) is documented here:
NetBackup in Highly Available Environments Administrator's Guide
11-28-2016 07:09 AM
Sorry - I do not think such a setup can be called HA.
This text is from Wikipedia definition of High Availability
There are three principles of systems design in reliability engineering which can help achieve high availability.
Elimination of single points of failure. This means adding redundancy to the system so that failure of a component does not mean failure of the entire system.
Reliable crossover. In redundant systems, the crossover point itself tends to become a single point of failure. Reliable systems must provide for reliable crossover.
Detection of failures as they occur. If the two principles above are observed, then a user may never see a failure. But the maintenance activity must.
As you can see, anything "single" will be fall for the first rule - "Elimination of single points of failure".
Talk to the customer about what types of failures customer want to protect against. if full DR is not possible due to lack of funds, then something "more" that a regular setup is enough. But customer need specify that "more".
11-28-2016 11:22 AM
Hello,
As you mentioned in your description I don’t think you can configure the HA for NBU.
If customer wants HA then set customer expectation about cost.
Add-on:
We can reduce the hardware cost by configure the standalone cluster and set the value for "OnlineRetryLimit" attribute for resources in VCS.
The agent monitor the resources and if there is any failure then it will try to start the failed resources on same system.
VCS:
https://sort.veritas.com/public/documents/sf/5.1/aix/html/vcs_admin/ap_vcs_attributes4.html
MS-Cluster
https://support.microsoft.com/en-us/kb/228923
All this things can be possible if you configure the NBU in cluster application.
Note: It won't protect from any hardware failure, OS failure.
Thanks & Regards,
11-29-2016 12:35 AM
Customer asked us to provide a solution for high availabilty if the VM(Master Server) went down but without vMotion and cluster enablement using supported technologies by Veritas Netbackup due to licensing issue.
On the other hand, we proposed him to perform cloning of a VM(Master Server) and /usr/openv will be from the shared disk on both Machines and cloned VM will be powered off. Which he denies to present the RDM Lun to VM.
While, As a virtualized application, NetBackup is unaware of being deployed within a VM, and support for such operations is provided by the hypervisor vendor.
Please suggest what can be done in such case
11-29-2016 12:44 AM
Please suggest what can be done in such case
Walk away.
Your customer is not serious about High Avalability.
Catalog backup is all they can rely on to protect the master server.
11-29-2016 01:18 AM