in reply to Re: Re (tilly) 2: Cron Question
in thread Cron Question
Then again I have to work with crontabs which multiple developers have edit privileges on. When you see something in there that doesn't belong, it is very nice to be able to say, "Who did that?" and get an answer.
Plus we already use CVS for our source-code. And working from system backups is a pain. With CVS any developer has that visibility in the same way they track anything else, without having to ask anyone for assistance. Without it all recoveries have to go to the sysadmin with the first question being, "Where do we keep the backups?" And then once you know where you keep the backups, "Where does this version of Unix keep its crontabs?" My sysadmin has more important things to do than babysit me through what should be a minor "oops". Rolling back something in a crontab may be a common need, but it needn't be a common request.
An incidental benefit for some people is that using revision control makes it easy to have multiple machines have the same crontab.
And finally, revision control is more convenient. For instance I am in the process of modifying some logic about what happens where in the nightly processes we run. It is very convenient that I can edit the crontab file as I am also editing the code for the processes that will run, and then check the associated changes into production together. If I had to make my source-code changes and then remember everything to change in production, then I would be more likely to make mistakes as I take things into production.
So while keeping backups is sufficient for a personal machine, given the choice I strongly prefer to use revision control. (And I would walk before working on a code-base without any kind of revision control in a shared environment. Revision control isn't very hard to set up, and I don't consider it optional. One vote to the first person who can figure out what famous project I would criticize on this issue...)
|
|---|