1

What's a good approach to mirroring a production environment's data for dev? We have one production server in place that mounts many smb shares which several scripts run against routinely.

We now have a separate server for development that we want to keep separate for testing. How do I get sample data from all those smb shares without copying them all? The dev server couldn't hold all that data so I'm looking for something that could routinely run and just copy the first X files out of each directory.

The goal is to have the dev server be "safe" and not mount those same shares during testing.

user2283020
  • 7
  • 1

1 Answers1

0

For a development environment I like to have:

  • Known good data
  • Known (constructed) bad data
  • Random sample of live data

What I mean by "constructed" is data that I have put together in a certain way so I know exactly how it is bad.

In your case I'd have my good and bad data and then write a little Bash script to copy data from the SMB shares to the local dev machine. Maybe run a ls -t on each of the shares so you can grab the newest files, save that output to a file and use head or some other utility to read the first N lines - and copy those files to your dev machine.

pseudo code

clear data directory
copy known good data from some local directory
copy known bad data from some local directory
begin loop: for every SMB share
   run `ls -t` and output the results to a file
   run `head` or some other utility to get the first N lines (ie file names)
   copy those files from the SMB share to my local data directory
end loop

You could set up cron to execute this little script however often you want.

Benny Hill
  • 6,191
  • 4
  • 39
  • 59
  • Not a bad bit of pseudocode, and good advice on known-good and bad. Something like that should work. It'll have to recurse into the directories, but that doesn't seem overly daunting. – user2283020 Apr 15 '13 at 17:40