Recently, I read this article on jails and vnets, listed below. I worked through the tutorials alright. What got my attention, though, was that there was nothing in the jails. Nothing. They're empty. Most of the time, I have worked exercises and tutorials that filled out the jails with a copy of the operating system.
REF: https://issue.freebsdfoundation.org/publication/?i=651491&ver=html5&p=19
So, I would like to ask: What's the least amount of files and programs that we need to provide for a jail to perform a task? That is, can a jail have just enough memory attached to run one process, like a socket listener program, and do that? What about a couple of processes? What about a jail with five or six processes running among a collection of about 20 files?
When do we need to prepare to attach a full copy of the operating system inside it? Do we see jails with copies of the OS in there because we frequently want to attain the goal of virtualizing the whole operating system; or, if we provide a small collection of executable programs in a jail, then will they fail to run without a separate copy of the OS? If we design jails to run programs, when do we get to the point that we begin to require a copy of the OS in there?
Just curious. I saw a great difference in the empty jails in that FreeBSD Journal article and other tutorials, like FreeBSD Mastery: Jails. Any help you might provide would be appreciated. Thanks.
REF: https://issue.freebsdfoundation.org/publication/?i=651491&ver=html5&p=19
So, I would like to ask: What's the least amount of files and programs that we need to provide for a jail to perform a task? That is, can a jail have just enough memory attached to run one process, like a socket listener program, and do that? What about a couple of processes? What about a jail with five or six processes running among a collection of about 20 files?
When do we need to prepare to attach a full copy of the operating system inside it? Do we see jails with copies of the OS in there because we frequently want to attain the goal of virtualizing the whole operating system; or, if we provide a small collection of executable programs in a jail, then will they fail to run without a separate copy of the OS? If we design jails to run programs, when do we get to the point that we begin to require a copy of the OS in there?
Just curious. I saw a great difference in the empty jails in that FreeBSD Journal article and other tutorials, like FreeBSD Mastery: Jails. Any help you might provide would be appreciated. Thanks.
