Date: Tue, 16 Dec 2014 19:09:50 +0000 From: Bruce Simpson <bms@fastmail.net> To: Bryan Venteicher <bryanv@FreeBSD.org>, src-committers@freebsd.org, svn-src-projects@freebsd.org Subject: Re: svn commit: r275835 - projects/paravirt Message-ID: <5490837E.8050907@fastmail.net> In-Reply-To: <201412161857.sBGIvgKR045243@svn.freebsd.org> References: <201412161857.sBGIvgKR045243@svn.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Bryan, I am very glad to see this, because: there are some limitations with Bhyve right now which make a KVM based workflow more suitable for my specific use case (protocol testbed on laptop). On 16/12/2014 18:57, Bryan Venteicher wrote: > This branch will initially be used to develop support for KVM > paravirtualization - KVM clock, EIO, and async page fault - > and then hopefully pvops infrastructure and support for other > hypervisors. > KVM is likely to be a better choice in constrained environments (e.g. laptops) because of the richer support for sparse storage (QCOW2, virtio-scsi, virtio-block and TRIM etc.) Unfortunately one limitation of this approach is that ZFS on Linux has not yet gained TRIM features, so this rules out direct migration to Bhyve on the same system at a later date; I will have to use ext4 for guests hosted on a laptop SSD and migrate everything manually. I have discussed the lack of sparse storage support in BHyve in private with marcel@. He is aware of the issue and is working to address it in tools, however it does need to be looked at further afield. I cannot get directly involved in this due to other commitments. thanks Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5490837E.8050907>