Re: [PATCH v4 00/12] virtio: s4 support
From: Amit Shah <hidden>
Date: 2011-12-08 11:56:50
Also in:
lkml
Possibly related (same subject, not in this thread)
- 2011-12-07 · Re: [PATCH v4 00/12] virtio: s4 support · Dor Laor <hidden>
- 2011-12-07 · Re: [PATCH v4 00/12] virtio: s4 support · Gleb Natapov <hidden>
- 2011-12-07 · Re: [PATCH v4 00/12] virtio: s4 support · Rusty Russell <hidden>
- 2011-12-06 · [PATCH v4 00/12] virtio: s4 support · Amit Shah <hidden>
On (Wed) 07 Dec 2011 [21:22:41], Rusty Russell wrote:
On Wed, 7 Dec 2011 13:14:56 +0530, Amit Shah [off-list ref] wrote:quoted
On (Wed) 07 Dec 2011 [17:54:29], Rusty Russell wrote:quoted
I figure there's a reason, but it seems a bit weird :)Well, there is one reason right now: migrating storage along with VMs. The guest needs to sync all data to the disk before the target host accesses the image file.I don't see why. Sure, if qemu (or whatever) is doing buffering, it needs to flush that. It doesn't, last I looked, it just maps to read/write. What am I missing?
I'm sorry, I got confused. The external file migration Gleb mentioned, though, is something that libvirt uses right now and causes guest time craziness so S4 is one (and natural) way to solve it. Amit