A CentOS VM crashed badly, buggering up the entire filesystem so now it’s not even recognised. Don’t suppose there’s any way to do some kind of fsck on it?

Well I mounted it, but it’s properly buggered.

Where’s the vdisk stored exactly? No snapshotting on your SAN?

It’s my laptop. I use VMs for setting up different environments for dev.

Managed to use vmware disk tools to mount it but couldn’t get anything out of it. I have a copy at home, just means I can’t do what I wanted to tonight.

Environments for Dev? You really need to get Vagrantup.com

I couldn’t imagine actually building virtual environments by hand these days.

Rachel

If the vm disk was damaged, I’d just have to type vagrant destroy and then vagrant up to be sorted again.

Hmm. I only need ONE environment for each product, so setting all that up would be far more difficult than simply installing each setup by hand. It’s only for my personal use, no-one else uses it.

Vagrant = docker?

EDIT not quite.

Well, the fact it now needs rebuilding kind of demonstrates that’s not exactly true… 😉

It’ll still be quicker!

Product installation is complicated enough without having to script it through some third party palaver. Too heavyweight for my purposes.

I have a backup in this case, as it happend, but as policy I don’t. I’d have to rebuild which means starting from a template I have and running the installers I have, for this particular environment it’s simple. But for others it’s totally different.

No, vagrant provides “control” over VMs. Allows you to define a fm in files, including how it should be configured etc. Then all you have to do is start it and magic happens. Out pops a perfect vm each time.

Means you can share a vm with someone simply by sending the definition which will then auto-build on their machine.

Allows the vm host sw to be either VMware or Virtualbox.

Means you can share a vm with someone simply by sending the definition which will then auto-build on their machine.

Still doesn’t seem beneficial for me, given how infrequently I have to duplicate or share anything.

Okay, if nothing else, it is always good practice to remove manual procedures from any development process; they are too easily prone to error. Often, the effort required to script or otherwise define the build means that the build is better thought out and more “robust”.

I’ll be surprised if each Dev environment you create is really that unique.

Rachel

You are thinking in terms of running a dev environment for a team, I think, which is not what I’m doing.

I’ll be surprised if each Dev environment you create is really that unique.

They all have different product stacks on, that’s why I have them!

Okay, this is starting to feel a little bit like a broken Volkswagen…

Can you boot a linux rescue CD in your VM, and then just run fsck manually from your rescue shell?

Or alternatively, there’s a command for turning a VM disk into a real physical disk you can mount in the host. I can’t remember what it’s called though, I’ll have a quick look.

EDIT: kpartx might help.