Date: Mon, 14 Nov 2022 21:25:36 +0000 From: bugzilla-noreply@freebsd.org To: virtualization@FreeBSD.org Subject: [Bug 267769] Bhyve core dump on suspend/resume of virtio-scsi device Message-ID: <bug-267769-27103-Hoqmi7Fyaz@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-267769-27103@https.bugs.freebsd.org/bugzilla/> References: <bug-267769-27103@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D267769 --- Comment #2 from dan@sunsaturn.com --- Well according to a lot of internet posts , virtio-scsi was designed to rep= lace virtio-blk. It is very convenient to pass in 10 devices to a guest with a 1 liner to bhyve with a cam ioctl device.=20 Also FreeBSD cannot fix vfs.zfs.vol.recursive because of the recursive dead lock issues, so only way currently to mount a ZFS guest to a host is using iscsi /dev/da* devices. While we can keep using virtio-blk devices the issue arises on new install = of guests such as linux guests for example where creating virtio-blk devices creates devices such as /dev/vda*. When passing virtio-scsi, the devices are /dev/sda*, so this complicates issues greatly. While FreeBSD guests do not suffer issues of being able to label GPT partit= ions in /etc/fstab, Linux guests hardcode all their GPT labels to physical disks like /dev/vda* which would make them unbootable switching from virtio-blk to virtio-scsi down the road. So I am not sure what is best course of action here, install all guests as virtio-scsi in preparation of suspend/resume functionality, or go with virtio-blk and have to reinstall all the guests at some point. Dan. --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-267769-27103-Hoqmi7Fyaz>