Date: Tue, 7 Aug 2018 21:38:42 +0200 From: Jakub Chromy <hicks@cgi.cz> To: "freebsd-virtua." <freebsd-virtualization@freebsd.org> Subject: Re: Seeking advice on virtualization Message-ID: <91bddabb-5fba-ad01-fb6f-063e673a0ba1@cgi.cz> In-Reply-To: <CADdqeiOeFn1F8cqWuVnskiU4Xa720HLJQcaN-CJ8Qbkrp%2BH2Tw@mail.gmail.com> References: <20180805224205.GB17784@tau1.ceti.pl> <CADdqeiM8rNGGgw7vyUT%2BCGxmcct93NK9xHCFGf4g2B6dG_xhrA@mail.gmail.com> <20180807170648.GA1599@tau1.ceti.pl> <a88a341c-fb3b-453e-7151-9e992f35586a@cgi.cz> <CADdqeiOeFn1F8cqWuVnskiU4Xa720HLJQcaN-CJ8Qbkrp%2BH2Tw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> even cooler you get 'clones' so you can say make a raw/zvol of debian > or whatever you like then if you make a snapshot of it and clone it to > something else 'debian2' it uses no space until you write/delete/edit > something, basically the clone only has diffs. yep... but it has some downsides too.. try to delete the original (source) ZVOL. Impossible until you remove the daughter clones first. > And yes zvols are literally like raw devices, dd style you can infact > take a raw image and dd it to a zvol yes... and you can even create sparse volumes (-s flag).. supercool :) > > On 7 August 2018 at 19:57, Jakub Chromy <hicks@cgi.cz > <mailto:hicks@cgi.cz>> wrote: > > I was writing about "hard disk file" format, in which a hypervisor > (i.e. bhyve, kvm, virtualbox) is keeping a disk for emulated > machine. Wikipedia calls it "img format": > > https://en.wikipedia.org/wiki/IMG_(file_format) > <https://en.wikipedia.org/wiki/IMG_%28file_format%29>; > > Advantage from using this format (as opposed to something like > qcow or > vmhd) is that, in theory (and even in practice) one can boot such > machine (I mean, virtual machine defined with such "raw" hard > drives) > using any hypervisor. > > > ZFS ZVOL is a true "raw device" as well... (or at least it did > behave like that for me): > > dd if=/dev/zvol/pool/mypornhubpremiumarchive0 > of=/var/vm/mypornhubpremiumarchive0.raw > > but you get snapshots, zfs send | zfs recv and stuff. > > -- > > > regards / s pozdravem > > > Jakub Chromy > > > CGI Systems div. > ---------------- > CGI CZ s.r.o. > sales@cgi.cz <mailto:sales@cgi.cz> > 775 144 257 > 234 697 102 > www.cgi.cz <http://www.cgi.cz>; > > > On 7.8.2018 19:06, Tomasz Rola wrote: > > On Mon, Aug 06, 2018 at 12:07:13AM +0100, Paul Webster wrote: > > In theory as ZFS works on both linux and BSD you could > simply use vdevs and > snapshots for easy transport > > Um-hm. > > I was writing about "hard disk file" format, in which a hypervisor > (i.e. bhyve, kvm, virtualbox) is keeping a disk for emulated > machine. Wikipedia calls it "img format": > > https://en.wikipedia.org/wiki/IMG_(file_format) > <https://en.wikipedia.org/wiki/IMG_%28file_format%29>; > > Advantage from using this format (as opposed to something like > qcow or > vmhd) is that, in theory (and even in practice) one can boot such > machine (I mean, virtual machine defined with such "raw" hard > drives) > using any hypervisor. Or to put it differently, it is not > proprietary > and is the easiest one to implement, so it is what most > probably will > keep being used years or decades from now (in whatever > hypervisor / PC > emulator of the future day is fashionable). > > I believe in the past I have installed an OS (say, FreeDOS) using > virtualbox and after deciding I would not use virtualbox in a > future, > I started to boot said machine using kvm. Likewise, I believe some > OSes rejected being installed under certain hypervisor, so one > had to > install them using this other hypervisor and then could happily > continue to run it under his preferred hypervisor. > > All of this made possible thanks to avoiding file formats > supported by > one or only few hypervisors. > > Of course there are many hd-file formats and some are supported by > more than one hypervisor, but the easiest one is raw and in > case of > emergency it can be also mounted as any other block device > (always, I > guess, but I would pay attention to block size mismatch). > > > _______________________________________________ > freebsd-virtualization@freebsd.org > <mailto:freebsd-virtualization@freebsd.org> mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-virtualization > <https://lists.freebsd.org/mailman/listinfo/freebsd-virtualization>; > To unsubscribe, send any mail to > "freebsd-virtualization-unsubscribe@freebsd.org > <mailto:freebsd-virtualization-unsubscribe@freebsd.org>" > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?91bddabb-5fba-ad01-fb6f-063e673a0ba1>