Date: Tue, 19 Jan 2016 10:36:45 -0800 From: Cy Schubert <cschuber@gmail.com> To: Shawn Webb <shawn.webb@hardenedbsd.org>, Alan Somers <asomers@freebsd.org> Cc: Kurt Lidl <lidl@pix.net>, "svn-src-head@freebsd.org" <svn-src-head@freebsd.org>, "svn-src-all@freebsd.org" <svn-src-all@freebsd.org>, "src-committers@freebsd.org" <src-committers@freebsd.org>, "cy@freebsd.org" <cy@freebsd.org> Subject: RE: svn commit: r294329 - inhead/sys/cddl/contrib/opensolaris/uts/common/fs/zfs: . sys Message-ID: <569e823b.cc0c620a.88044.ffffd719@mx.google.com>
next in thread | raw e-mail | index | archive | help
Just another thought. If we must do this, why not a sysctl to restore funct= ionality to those of us who need it. If indeed there are panics and hangs d= ue to zpools on vdevs, a sysctl to bypass the restriction, in effect saying= "I understand the risk," would be a reasonable compromise. I'm willing to = put together a patch to do this. Sent from my cellphone, ~Cy -----Original Message----- From: Cy Schubert Sent: 19/01/2016 10:27 To: Shawn Webb; Alan Somers Cc: Kurt Lidl; svn-src-head@freebsd.org; svn-src-all@freebsd.org; src-commi= tters@freebsd.org; cy@freebsd.org Subject: RE: svn commit: r294329 - inhead/sys/cddl/contrib/opensolaris/uts/= common/fs/zfs: . sys Agreed, I do a fair bit of testing (nothing permanent) by putting zpools on= zvols. I have never experienced a hang or panic due to zpools on zvols. Th= is is a POLA violation. Regarding zfs hangs, I only experience them with zpool vdevs on USB devices= on my laptop when I forget to zpool export tpool prior to laptop shutdown/= reboot. This is acceptable as I'm in front of it. Sent from my cellphone, ~Cy -----Original Message----- From: Shawn Webb Sent: 19/01/2016 10:07 To: Alan Somers Cc: Kurt Lidl; svn-src-head@freebsd.org; svn-src-all@freebsd.org; src-commi= tters@freebsd.org Subject: Re: svn commit: r294329 - inhead/sys/cddl/contrib/opensolaris/uts/= common/fs/zfs: . sys On Tue, Jan 19, 2016 at 10:20:38AM -0700, Alan Somers wrote: > On Tue, Jan 19, 2016 at 10:08 AM, Kurt Lidl <lidl@pix.net> wrote: > > Removing the ability to run different zpools on top of a zvol is > > a significant reduction in functionality of the entire system, and a hu= ge > > violation of the POLA. >=20 > The thing is, it never really worked in the first place. Panics and > deadlocks are so frequent that I don't think the feature was usable > for anybody. I actively use this every day and will be making even more use of it in the very near future. I haven't had a single kernel panic. I have had deadlocks at shutdown, but since it's a laptop and I'm sitting right in front of it, it's not a big deal. I'd say fix the underlying problem, don't prevent people from getting work done. >=20 > > > > At the very least, can you not add a sysctl that allows the > > dangerous behavior (even if it defaults to off)? Myself > > and certainly others rely on having being able to use a zpool > > installed into a zvol for hosting bhyve virtual machines. >=20 > Your use case should be unaffected. The guest has a different ZFS > instance than the host, so it should work just fine. Please let me > know if you have problems. --=20 Shawn Webb HardenedBSD GPG Key ID: 0x6A84658F52456EEE GPG Key Fingerprint: 2ABA B6BD EF6A F486 BE89 3D9E 6A84 658F 5245 6EEE
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?569e823b.cc0c620a.88044.ffffd719>