I have been trying to come up with a way to create as nearly a one stop shopping experience as possible with WFA. After much trial and error, command swapping, and workflow rebuilds I wanted to put this forward for others to possibly use as a template for their own needs.
A quick idea on our environment:
We are entirely 7 mode and this workflow is only for 7 mode
We have multiple OCUM instances running Protection Manager
We have frames at 5 major sites and multiple smaller sites so static information is not practical, all data must be as dynamically updated by OCUM/PM as possible
For NFS we have four default allocation possibilities:
- Standalone
- standalone with DR
- in a vfiler
- in a vfiler with DR.
On top of that we have three default volume allocation types
- snapvault backup - options nosnap=on set and snap sched 0 0 0, snap reserve 20%
- Must be put into specific PM dataset based off of the frame name
- local snaps - options nosnap=off snap sched 2 7 0, snap reserve 20%
- no snaps - options nosnap=on snap sched 0 0 0, snap reserve 0%
But we have allocations which require 50% snap reserve
Using dependencies, commands executing only if variable are set to true or false, any of these allocations can be done with the same workflow. Note that there are commands that require PERL to be installed because I am not allowed under any circumstances to run an OpenSSH server on a production system. There are certified powershell commands that you can substitute for everything except the reporting group add, that work great.
By toggling the check box for vfiler, it enables the vfiler query boxes, which are pulled from OCUM and only list the vfilers available on that frame and enables the "add volume to vfiler" and "create vfiler export" commands
By toggling the check box for DR, it enables the workflow DR portions of the system
Using enumeration boxes I am able to provide only the reporting groups that are relevant as they are much less dynamic than vfiler/volumes
Using SQL queries I am able to lock down a great many things so that the workflow is more "usable" by less skilled users.
You can lock a good many things down farther, however we found that it was locked down too much and had to be opened to this point as it requires extremely good planning in the layout of the dataset names.
There are still some things to be cleaned up in this workflow, but I hope it can help others to look at WFA and what it can do in a different light. By setting down and rebuilding my workflows by looking at them from different angles, I have gone from sixteen workflows around CIFS and NFS down to two.
Any comments on how I can do any of this better are greatly appreciated. Learning by trial and error is great, but you don't tend to learn the best way.