Bug 1294884

Summary:	Support for Virt-FS (via NFS + virtio-vsock) (libvirt)
Product:	Red Hat Enterprise Linux 7	Reporter:	Ademar Reis <areis>
Component:	libvirt	Assignee:	Ján Tomko <jtomko>
Status:	CLOSED DEFERRED	QA Contact:	yafu <yafu>
Severity:	high	Docs Contact:
Priority:	high
Version:	7.3	CC:	areis, bugproxy, dyuan, fs-qe, hannsj_uhl, hhan, jomiller, jsuchane, jtomko, juzhang, mtessun, mzhan, stefanha, steved, xuzhang, yafu, yoyang
Target Milestone:	rc	Keywords:	FutureFeature
Target Release:	7.7
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:	1294879
Clones:	1518997 (view as bug list)		Environment:
Last Closed:	2018-12-03 16:35:54 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:	1291282, 1291284, 1291286, 1291851, 1294879, 1294880, 1315822, 1378137, 1382695, 1518995, 1518996
Bug Blocks:	1363787, 1518997, 1553272, 1563290, 1577865

Description Ademar Reis 2015-12-30 19:46:06 UTC

Once we have support for virtio-sock everywhere (QEMU, Kernel and NFS), libvirt will have to provide some mechanism for a Virt-FS feature, where a folder gets shared between host and guest.

+++ This bug was initially created as a clone of Bug #1294879 +++

We'we working on virtio-vsock in QEMU/KVM, which implements a host/guest communication device, allowing applications from both sides to communicate without the need of a network. It implements the POSIX Sockets API and supports both SOCK_STREAM and SOCK_DGRAM.

For virtio-vsock support in QEMU/kernel, see bug 1291282 and bug 1291284.

Once virtio-vsock is implemented, we want NFS (both server and client) to support it so that we can use it to share directories between the host and the guest, a feature we call "Virt-FS".

Comment 2 Stefan Hajnoczi 2016-09-21 15:05:52 UTC

*** Bug 1338715 has been marked as a duplicate of this bug. ***

Comment 4 IBM Bug Proxy 2017-06-19 19:21:20 UTC

------- Comment From lagarcia.com 2017-06-19 15:17 EDT-------
Is this on track for RHEL 7.4?

Comment 5 Ademar Reis 2017-06-19 23:41:54 UTC

(In reply to IBM Bug Proxy from comment #4)
> ------- Comment From lagarcia.com 2017-06-19 15:17 EDT-------
> Is this on track for RHEL 7.4?

Unfortunately not. We're still working on upstream support for virtio-vsock in NFS and the complete feature has been deferred to a future release of RHEL or layered products.

Comment 11 IBM Bug Proxy 2018-03-23 14:21:25 UTC

------- Comment From lagarcia.com 2018-03-23 10:20 EDT-------
Ademar,

Is this on track for RHEL 7.5?

Comment 12 Ademar Reis 2018-03-23 15:00:21 UTC

(In reply to IBM Bug Proxy from comment #11)
> ------- Comment From lagarcia.com 2018-03-23 10:20 EDT-------
> Ademar,
> 
> Is this on track for RHEL 7.5?

Unfortunately not. We couldn't get upstream NFS to accept vsock support yet, so this is still in progress. We had to submit a IETF RFC to get it accepted as a standard and that process takes time. Even RHEL-7.6 is a stretch target.

Comment 13 Ademar Reis 2018-12-03 16:35:54 UTC

Will not be implemented in RHEL-7.

For sometime in RHEL-8, we're planning a different approach, using virtio-fs. This is being tracked via Bug 1519458 and Bug 1519459.

Comment 14 IBM Bug Proxy 2018-12-04 08:00:25 UTC

------- Comment From KURZGREG.com 2018-12-04 02:56 EDT-------
(In reply to comment #16)
> Will not be implemented in RHEL-7.
> For sometime in RHEL-8, we're planning a different approach, using
> virtio-fs. This is being tracked via Bug 1519458 and Bug 1519459.

Access to these BZs is denied, and they don't seem to have been mirrored to IBM's BZ.

Please provide access. My login in RH's BZ is gkurz.ibm.com.

Comment 15 Ján Tomko 2018-12-04 14:27:05 UTC

I made bug 1519459 public since it does not contain any different information than this one.

Comment 16 Ademar Reis 2018-12-04 17:01:50 UTC

(In reply to Ján Tomko from comment #15)
> I made bug 1519459 public since it does not contain any different
> information than this one.

Ditto for Bug 1519458.