Answer the question
In order to leave comments, you need to log in
What are the options for reinstalling the OS on a server in production?
Colleagues, welcome! And immediately the question is: how possible (or what nuances can be) is such a conditional scenario, in which the system is first installed on a virtual machine and deployed using (from a familiar, for example) VirtIO drivers, then (for example) clonezill is transferred to the system server disk - and so that everything works right away or a little later. For, as I understand it, there may be problems with drivers that (in theory) when connected to the network are pulled by the system during installation (like the same Ubuntu)? Or it is worth using, perhaps, the processes and (or) platforms known to you for deploying to production (also, I would like to clarify, if possible, how to organize the infrastructure for this), maybe something like Docker. Or everything is much simpler and you can get by with updating and upgrading with a package manager after deployment, or you can put drivers manually or configure native ones for the machine in advance (it’s also a question of how to organize this process correctly). Or is it inappropriate initially, at the stage of conception, and is it easier to raise a server, quickly migrate virtual machines there and do business on the spot?
Answer the question
In order to leave comments, you need to log in
It's about a Linux server. Drivers in the kernel (if we are not talking about printers, etc.). Also, we are not talking about hardware RAID (because mdadm). Most likely it will work right away. Clonezilla is not needed, dd is enough.
But the word "production" is annoying, because. it would be better to pre-train, do all this on test hardware. And be sure to provide a "plan B" - a quick rollback of changes if something goes wrong.
Added to the section of similar questions (the task is different there, but the meanings intersect): How to raise a Linux server on VirtualBox with the possibility of subsequent transfer to a physical server?
I did not understand the situation in which this is all required.
Ideally, an operating system is installed on the server, which acts as a host; or even an operating system designed to be Dom0. It is not reinstalled unless absolutely necessary.
And inside virtual machines, images of the necessary virtual machines are already uploaded.
PS: It seems that all drivers should. included in the operating system; and even hard-coded into the kernel. By kr.mere, in FreeBSD - exactly so.
It’s easier to set up a server, quickly migrate virtual machines there and do business on the spot
I use the option with such cloning only for a bunch of client PCs (miniPCs) with ODA settings and a set of software. The computers are all variegated, but clonezilla copes, rules and fstab itself and other necessary things.
In terms of the server, we deploy the mini-iso with our hands, then we configure it for the task (with our hands, scripts, if often, a lot), then we already transfer the information there and the services. After the end of the sync raid)
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question