Skip to main content
Answer

Pre/Post process - anyone tried setting up a check to see if EC2 is running first

  • May 31, 2022
  • 7 replies
  • 91 views

Forum|alt.badge.img+1

Hello - 

We have a bunch of DEV servers that the developers turn on and off without letting us know. We would like to setup either a prescan or prebackup process that would check to see if the server is active, either by a network check or seeing if its powered on  - and if its not, skip the backup. 

Anyone tried this?  

thanks

Paul

Best answer by christopherlecky

You should probably do it with a workflow

7 replies

christopherlecky
Byte
Forum|alt.badge.img+17

You should probably do it with a workflow


Forum|alt.badge.img+1
  • Author
  • Bit
  • May 31, 2022

after I posted this I started poking around and looks like workflow would be the best course for doing what I need.

thanks


Onno van den Berg
Commvault Certified Expert
Forum|alt.badge.img+20

Those instances are EC2 instances, right? Why not leverage VSA backup? That way the powerstate becomes irrelevant and you can continue creating backups without a need for a worklow or script. 


Forum|alt.badge.img+1
  • Author
  • Bit
  • May 31, 2022

Those instances are EC2 instances, right? Why not leverage VSA backup? That way the powerstate becomes irrelevant and you can continue creating backups without a need for a worklow or script. 

We do use VSA for a few of our bigger EC2’s - but have decided due to cost to stick with the old file system backup strategy with fewer VSA’s. 


Onno van den Berg
Commvault Certified Expert
Forum|alt.badge.img+20

What kind of cost are you talking about if I may ask? Do you keep the DEV backups for a long time? What are protecting now with the agent as I assume you are protecting the entire VM because you most likely do not have context with the stuff that they are doing. You already have access nodes in place so in case of a streaming backup it's just a matter of some processing cost and some cost for storage for the unique blocks. Being able to deliver back a broken dev box to a developer might be a big win as well as this will also save cost ;-) 


Forum|alt.badge.img+1
  • Author
  • Bit
  • May 31, 2022

mainly the additional proxies we would need is what is driving the cost. About 25 + more would be needed for our footprint. For prod, we have infinite retention which in turn eats our budget so we have to do everything else as cost effective as possible.  


Onno van den Berg
Commvault Certified Expert
Forum|alt.badge.img+20

So it's a big environment, right? How many VM's are in scope and again what's te retention? In case you want to keep them for a short period of time, say 7 days, then you could consider using snapshots alone. That would also result in cost, but as snapshots re incremental combined with a short retention it might be something to consider. 

B.t.w. are you leveraging the scale-out capabilities of Commvault? 

Oh, and b.t.w. about the way to automate it yourself. I would stay away from workloads and just create some script leveraging the APIs to handle this. It seems you have a lot of dev guys around so plenty of help in the neighborhood ;-)