Date: Fri, 10 Nov 2023 23:47:23 +0100 From: Martin Matuska <mm@FreeBSD.org> To: Xin LI <delphij@gmail.com> Cc: d@delphij.net, FreeBSD-STABLE Mailing List <freebsd-stable@freebsd.org>, pjd@freebsd.org Subject: Re: Is 14.0 to released based on 0 for sysctl vfs.zfs.bclone_enabled ? Message-ID: <77ad9593-34a8-48dc-8533-aafb852f1d19@FreeBSD.org> In-Reply-To: <CAGMYy3tqwa_JQ01BKLVfSKt9N%2BWyK=M13j_i3kT=LJ_-MQyrQQ@mail.gmail.com> References: <2F81D978-7DBD-42CE-8ECF-C020B0CB5C29.ref@yahoo.com> <2F81D978-7DBD-42CE-8ECF-C020B0CB5C29@yahoo.com> <7a906956-6836-421e-b25e-ff701369e3ed@FreeBSD.org> <BBFDD30F-FB5D-44C8-ADA7-5B5AF859D86A@karels.net> <830CD3A8-DB62-418D-A7F7-8DA6CB46B1F5@yahoo.com> <05b493bc-94a5-4c78-bebf-5581addc5b7b@FreeBSD.org> <47c5b902-eea6-4194-b84a-99a6343f6bd0@delphij.net> <ba2e7bdc-68ba-4093-816a-2f0ea5bb6a07@FreeBSD.org> <CAGMYy3tqwa_JQ01BKLVfSKt9N%2BWyK=M13j_i3kT=LJ_-MQyrQQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format. --------------hcDCO5UsWNX1hkTPj6uApmeT Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit The issues I had to deal with went away by deleting the problematic files (for good, no snapshot copies left). Deleting a dataset should be even better. On 10. 11. 2023 17:58, Xin LI wrote: > > > On Fri, Nov 10, 2023 at 7:50 AM Martin Matuska <mm@freebsd.org> wrote: > > Hi Xin, > > since when have you been using block cloning on the system? Is it > possible that there is already corrupted block-cloned data from the > > > That's a good question, I can't 100% rule out this possibility. I was > following -CURRENT in ~weekly to ~monthly on that system, and the pool > was created in March 2014. > > Do you think I should try rebuilding the pool from scratch? I do have > remote backup on a different server but was avoiding it because it's > time consuming. > > past? Is everything on one dataset or are you using multiple datasets > for /usr/src and /usr/obj? > > > /usr/src and /usr/obj are separate datasets, and the system runs > Poudriere so it have multiple copies of slightly different /usr/src > and /usr/obj's. > > Is there a way to identify datasets with block cloning, by the way? > Maybe I should try recreating these datasets first? > > > > Best regards, > mm > > On 10. 11. 2023 8:04, Xin Li wrote: > > On 2023-11-05 16:34, Martin Matuska wrote: > >> OpenZFS 2.2.0 in FreeBSD 14 fully supports block cloning. You can > >> work with pools that have feature@block_cloning enabled. > >> The sysctl variable vfs.zfs.bclone_enabled affects the behavior of > >> zfs_clone_range() which is called by copy_file_range(). When it is > >> set to 0, zfs_clone_range() does not do block cloning. > >> If it is set to anything else than 0, zfs_clone_range() does block > >> cloning (if all conditions are met - same ZFS pool, correct data > >> alignment, etc.). > >> > >> In FreeBSD-main, this tunable is enabled and I plan to enable > it in > >> stable/14 somewhere around December 11, 2023. > >> > >> As of today I personally use block cloning on all my systems. > > > > I'd like to share a different data point. It still panics on my > > storage (running -CURRENT about a week ago) when enabled and can be > > triggered by "make buildworld buildkernel". I wasn't able to > capture > > earlier coredump until the most recent one, which panicked with: > > > > > > cpuid = 2 > > time = 1699593456 > > KDB: stack backtrace: > > db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame > > 0xfffffe022f2bd7e0 > > vpanic() at vpanic+0x132/frame 0xfffffe022f2bd910 > > spl_panic() at spl_panic+0x3a/frame 0xfffffe022f2bd970 > > dmu_brt_clone() at dmu_brt_clone+0x555/frame 0xfffffe022f2bd9e0 > > zfs_clone_range() at zfs_clone_range+0xa4c/frame 0xfffffe022f2bdbb0 > > zfs_freebsd_copy_file_range() at > > zfs_freebsd_copy_file_range+0x18a/frame 0xfffffe022f2bdc30 > > vn_copy_file_range() at vn_copy_file_range+0x163/frame > 0xfffffe022f2bdce0 > > kern_copy_file_range() at kern_copy_file_range+0x380/frame > > 0xfffffe022f2bddb0 > > sys_copy_file_range() at sys_copy_file_range+0x78/frame > > 0xfffffe022f2bde00 > > amd64_syscall() at amd64_syscall+0x153/frame 0xfffffe022f2bdf30 > > fast_syscall_common() at fast_syscall_common+0xf8/frame > > 0xfffffe022f2bdf30 > > --- syscall (569, FreeBSD ELF64, copy_file_range), rip = > > 0x7fbb2da4ada, rsp = 0x7fbb02c5d48, rbp = 0x7fbb02c61e0 --- > > Uptime: 2h32m27s > > Dumping 7800 out of 32696 > > MB:..1%..11%..21%..31%..41%..51%..61%..71%..81%..91% > > > > #0 __curthread () at /usr/src/sys/amd64/include/pcpu_aux.h:57 > > #1 doadump (textdump=textdump@entry=1) at > > /usr/src/sys/kern/kern_shutdown.c:405 > > #2 0xffffffff80694480 in kern_reboot (howto=260) at > > /usr/src/sys/kern/kern_shutdown.c:526 > > #3 0xffffffff8069497f in vpanic (fmt=0xffffffff82603415 > "VERIFY3(nbps > > == numbufs) failed (%llu == %llu)\n", > ap=ap@entry=0xfffffe022f2bd950) > > at /usr/src/sys/kern/kern_shutdown.c:970 > > #4 0xffffffff8232999a in spl_panic (file=<optimized out>, > > func=<optimized out>, line=<unavailable>, fmt=<unavailable>) at > > /usr/src/sys/contrib/openzfs/module/os/freebsd/spl/spl_misc.c:103 > > #5 0xffffffff823a6605 in dmu_brt_clone > > (os=os@entry=0xfffff800c5ce4000, object=<optimized out>, > > offset=offset@entry=0, length=length@entry=207477, > > tx=tx@entry=0xfffff8071a108d00, bps=bps@entry=0xfffffe01e218c000, > > nbps=2, replay=0) > > at /usr/src/sys/contrib/openzfs/module/zfs/dmu.c:2303 > > #6 0xffffffff8250f67c in zfs_clone_range (inzp=0xfffff804416ac000, > > inoffp=0xfffff800b81cb048, outzp=0xfffff806f58f03a0, > > outoffp=0xfffff800b8063048, lenp=lenp@entry=0xfffffe022f2bdbf0, > > cr=0xfffff8000a6fe600) > > at /usr/src/sys/contrib/openzfs/module/zfs/zfs_vnops.c:1326 > > #7 0xffffffff8234b3ba in zfs_freebsd_copy_file_range > > (ap=0xfffffe022f2bdc48) at > > > /usr/src/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c:6294 > > #8 0xffffffff8079f443 in VOP_COPY_FILE_RANGE > > (invp=0xfffff804416cb1c0, inoffp=0xfffff800b81cb048, > > outvp=0xfffff806f51d3380, outoffp=0xfffff800b8063048, > > lenp=0xfffffe022f2bdd48, incred=0xfffff8000a6fe600, > flags=<optimized > > out>, > > outcred=<optimized out>, fsizetd=<optimized out>) at > > ./vnode_if.h:2385 > > #9 vn_copy_file_range (invp=invp@entry=0xfffff804416cb1c0, > > inoffp=inoffp@entry=0xfffff800b81cb048, > > outvp=outvp@entry=0xfffff806f51d3380, > > outoffp=outoffp@entry=0xfffff800b8063048, > > lenp=lenp@entry=0xfffffe022f2bdd48, flags=flags@entry=0, > > incred=0xfffff8000a6fe600, outcred=0xfffff8000a6fe600, > > fsize_td=0xfffffe022925b3a0) at /usr/src/sys/kern/vfs_vnops.c:3087 > > #10 0xffffffff8079a070 in kern_copy_file_range > > (td=td@entry=0xfffffe022925b3a0, infd=<optimized out>, > > inoffp=0xfffff800b81cb048, inoffp@entry=0x0, outfd=<optimized out>, > > outoffp=0xfffff800b8063048, outoffp@entry=0x0, > len=9223372036854775807, > > flags=0) at /usr/src/sys/kern/vfs_syscalls.c:4973 > > #11 0xffffffff8079a178 in sys_copy_file_range > (td=0xfffffe022925b3a0, > > uap=0xfffffe022925b7a0) at /usr/src/sys/kern/vfs_syscalls.c:5011 > > #12 0xffffffff80a97aa3 in syscallenter (td=0xfffffe022925b3a0) at > > /usr/src/sys/amd64/amd64/../../kern/subr_syscall.c:188 > > #13 amd64_syscall (td=0xfffffe022925b3a0, traced=0) at > > /usr/src/sys/amd64/amd64/trap.c:1194 > > #14 <signal handler called> > > #15 0x000007fbb2da4ada in ?? () > > > > > > and disabling bclone does appear to allow me to finish buildworld / > > buildkernel. > > > > The pool didn't have redaction_list_spill enabled. > > > > The ASSERT3U(nbps, ==, numbufs); in dmu_brt_clone was added when > block > > clone is first implemented. > > > > It seems that I am the only person who is seeing this as of > today. It > > seems that block clone was indeed being used for some data: > > > > saturn bcloneused 1.18M - > > saturn bclonesaved 1.21M - > > saturn bcloneratio 2.02x - > > > > The pool have dedup enabled for some datasets. > > > > Any suggestions? (In extreme cases I can recreate the storage pool > > from backup or copy the data somewhere else, then recreate the > pool, > > then copy data back, but I'd like to avoid that if possible) > > > > Cheers, > > > > > >> > >> mm > >> > >> On 04/11/2023 13:35, Mark Millard wrote: > >>> On Nov 4, 2023, at 04:38, Mike Karels <mike@karels.net> wrote: > >>> > >>>> On 4 Nov 2023, at 4:01, Ronald Klop wrote: > >>>> > >>>>> On 11/4/23 02:39, Mark Millard wrote: > >>>>>> It looks to me like releng/14.0 (as of 14.0-RC4) still has: > >>>>>> > >>>>>> int zfs_bclone_enabled; > >>>>>> SYSCTL_INT(_vfs_zfs, OID_AUTO, bclone_enabled, CTLFLAG_RWTUN, > >>>>>> &zfs_bclone_enabled, 0, "Enable block cloning"); > >>>>>> > >>>>>> leaving block cloning effectively disabled by default, no > >>>>>> matter what the pool has enabled. > >>>>>> > >>>>>> https://www.freebsd.org/releases/14.0R/relnotes/ also reports: > >>>>>> > >>>>>> QUOTE > >>>>>> OpenZFS has been upgraded to version 2.2. New features include: > >>>>>> • > >>>>>> block cloning, which allows shallow copies of blocks in file > >>>>>> copies. This is optional, and disabled by default; it can be > >>>>>> enabled with sysctl vfs.zfs.bclone_enabled=1. > >>>>>> END QUOTE > >>>>>> > >>>>> > >>>>> I think this answers your question in the subject. > >>>> I think so too (and I wrote that text). > >>> Thanks for the confirmation of the final intent. > >>> > >>> I believe this makes: > >>> > >>> QUOTE > >>> author Brian Behlendorf <behlendorf1@llnl.gov> 2023-05-25 > 20:53:08 > >>> +0000 > >>> committer GitHub <noreply@github.com> 2023-05-25 20:53:08 +0000 > >>> commit 91a2325c4a0fbe01d0bf212e44fa9d85017837ce (patch) > >>> tree dd01dfce6aeef357ade1775acf18aade535c6271 > >>> . . . > >>> Update compatibility.d files > >>> > >>> Add an openzfs-2.2 compatibility file for the next release. > Edon-R > >>> support has been enabled for FreeBSD removing the need for > different > >>> FreeBSD and Linux files. Symlinks for the -linux and -freebsd > names > >>> are created for any scripts expecting that convention. > Additionally, > >>> a symlink for ubunutu-22.04 was added. Signed-off-by: Brian > >>> Behlendorf <behlendorf1@llnl.gov> Closes #14833 > >>> END QUOTE > >>> > >>> technically incorrect in that compatibility.d/openzfs-2.2-freebsd > >>> should be distinct in content from compatibility.d/openzfs-2.2 so > >>> that block cloning would not be enabled. > >>> > >>> > >>>>>> Just curiousity on my part about the default completeness of > >>>>>> openzfs-2.2 support, not an objection either way. > >>>>>> > >>>>> > >>>>> I haven't seen new issues with block cloning in the last few > weeks > >>>>> mentioned on the mailing lists. All known issues are fixed > AFAIK. > >>>>> But I can imagine that the risk+effect ratio of data > corruption is > >>>>> seen as a bit too high for a 14.0 release for this particular > >>>>> feature. That does not diminish the rest of the completeness of > >>>>> openzfs-2.2. > >>>>> > >>>>> NB: I'm not involved in developing openzfs or the decision > making > >>>>> in the release. Just repeating what I read on the lists. > >>>> There was another block cloning fix in 14.0-RC4; see the > commit log. > >>>> Maybe there will be no more issues, but it seems that corner > cases > >>>> were > >>>> still being found recently. > >>> Looks like I'll stay at openzfs-2.1 pool features until there is > >>> a release that no longer has the default status: > >>> > >>> 0 for sysctl vfs.zfs.bclone_enabled > >>> > >>> I use main [so: 15 now] but only enable openzfs-2.* pool features > >>> supported by default on some FreeBSD release, that has an accurate > >>> compatibility.d/openzfs-2.*-freebsd file. > >>> > >>> === > >>> Mark Millard > >>> marklmi at yahoo.com <http://yahoo.com>; > >>> > >>> > >> > > > --------------hcDCO5UsWNX1hkTPj6uApmeT Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: 8bit <!DOCTYPE html> <html> <head> <meta http-equiv="Content-Type" content="text/html; charset=UTF-8"> </head> <body> <p>The issues I had to deal with went away by deleting the problematic files (for good, no snapshot copies left). Deleting a dataset should be even better.<br> </p> <div class="moz-cite-prefix">On 10. 11. 2023 17:58, Xin LI wrote:<br> </div> <blockquote type="cite" cite="mid:CAGMYy3tqwa_JQ01BKLVfSKt9N+WyK=M13j_i3kT=LJ_-MQyrQQ@mail.gmail.com"> <meta http-equiv="content-type" content="text/html; charset=UTF-8"> <div dir="ltr"> <div dir="ltr"> <div class="gmail_default" style="font-family:monospace,monospace"><br> </div> </div> <br> <div class="gmail_quote"> <div dir="ltr" class="gmail_attr">On Fri, Nov 10, 2023 at 7:50 AM Martin Matuska <<a href="mailto:mm@freebsd.org" moz-do-not-send="true" class="moz-txt-link-freetext">mm@freebsd.org</a>> wrote:<br> </div> <blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Hi Xin,<br> <br> since when have you been using block cloning on the system? Is it <br> possible that there is already corrupted block-cloned data from the</blockquote> <div><br> </div> <div> <div class="gmail_default" style="font-family:monospace,monospace">That's a good question, I can't 100% rule out this possibility. I was following -CURRENT in ~weekly to ~monthly on that system, and the pool was created in March 2014.</div> </div> <div class="gmail_default" style="font-family:monospace,monospace"><br> </div> <div class="gmail_default" style="font-family:monospace,monospace">Do you think I should try rebuilding the pool from scratch? I do have remote backup on a different server but was avoiding it because it's time consuming.</div> <div> </div> <blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">past? Is everything on one dataset or are you using multiple datasets <br> for /usr/src and /usr/obj?<br> </blockquote> <div><br> </div> <div> <div class="gmail_default" style="font-family:monospace,monospace">/usr/src and /usr/obj are separate datasets, and the system runs Poudriere so it have multiple copies of slightly different /usr/src and /usr/obj's.</div> <div class="gmail_default" style="font-family:monospace,monospace"><br> </div> <div class="gmail_default" style="font-family:monospace,monospace">Is there a way to identify datasets with block cloning, by the way? Maybe I should try recreating these datasets first?</div> <br> </div> <div> <div class="gmail_default" style="font-family:monospace,monospace"><span style="font-family:Arial,Helvetica,sans-serif"> </span><br> </div> </div> <blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> <br> Best regards,<br> mm<br> <br> On 10. 11. 2023 8:04, Xin Li wrote:<br> > On 2023-11-05 16:34, Martin Matuska wrote:<br> >> OpenZFS 2.2.0 in FreeBSD 14 fully supports block cloning. You can <br> >> work with pools that have feature@block_cloning enabled.<br> >> The sysctl variable vfs.zfs.bclone_enabled affects the behavior of <br> >> zfs_clone_range() which is called by copy_file_range(). When it is <br> >> set to 0, zfs_clone_range() does not do block cloning.<br> >> If it is set to anything else than 0, zfs_clone_range() does block <br> >> cloning (if all conditions are met - same ZFS pool, correct data <br> >> alignment, etc.).<br> >><br> >> In FreeBSD-main, this tunable is enabled and I plan to enable it in <br> >> stable/14 somewhere around December 11, 2023.<br> >><br> >> As of today I personally use block cloning on all my systems.<br> ><br> > I'd like to share a different data point. It still panics on my <br> > storage (running -CURRENT about a week ago) when enabled and can be <br> > triggered by "make buildworld buildkernel". I wasn't able to capture <br> > earlier coredump until the most recent one, which panicked with:<br> ><br> ><br> > cpuid = 2<br> > time = 1699593456<br> > KDB: stack backtrace:<br> > db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame <br> > 0xfffffe022f2bd7e0<br> > vpanic() at vpanic+0x132/frame 0xfffffe022f2bd910<br> > spl_panic() at spl_panic+0x3a/frame 0xfffffe022f2bd970<br> > dmu_brt_clone() at dmu_brt_clone+0x555/frame 0xfffffe022f2bd9e0<br> > zfs_clone_range() at zfs_clone_range+0xa4c/frame 0xfffffe022f2bdbb0<br> > zfs_freebsd_copy_file_range() at <br> > zfs_freebsd_copy_file_range+0x18a/frame 0xfffffe022f2bdc30<br> > vn_copy_file_range() at vn_copy_file_range+0x163/frame 0xfffffe022f2bdce0<br> > kern_copy_file_range() at kern_copy_file_range+0x380/frame <br> > 0xfffffe022f2bddb0<br> > sys_copy_file_range() at sys_copy_file_range+0x78/frame <br> > 0xfffffe022f2bde00<br> > amd64_syscall() at amd64_syscall+0x153/frame 0xfffffe022f2bdf30<br> > fast_syscall_common() at fast_syscall_common+0xf8/frame <br> > 0xfffffe022f2bdf30<br> > --- syscall (569, FreeBSD ELF64, copy_file_range), rip = <br> > 0x7fbb2da4ada, rsp = 0x7fbb02c5d48, rbp = 0x7fbb02c61e0 ---<br> > Uptime: 2h32m27s<br> > Dumping 7800 out of 32696 <br> > MB:..1%..11%..21%..31%..41%..51%..61%..71%..81%..91%<br> ><br> > #0 __curthread () at /usr/src/sys/amd64/include/pcpu_aux.h:57<br> > #1 doadump (textdump=textdump@entry=1) at <br> > /usr/src/sys/kern/kern_shutdown.c:405<br> > #2 0xffffffff80694480 in kern_reboot (howto=260) at <br> > /usr/src/sys/kern/kern_shutdown.c:526<br> > #3 0xffffffff8069497f in vpanic (fmt=0xffffffff82603415 "VERIFY3(nbps <br> > == numbufs) failed (%llu == %llu)\n", ap=ap@entry=0xfffffe022f2bd950) <br> > at /usr/src/sys/kern/kern_shutdown.c:970<br> > #4 0xffffffff8232999a in spl_panic (file=<optimized out>, <br> > func=<optimized out>, line=<unavailable>, fmt=<unavailable>) at <br> > /usr/src/sys/contrib/openzfs/module/os/freebsd/spl/spl_misc.c:103<br> > #5 0xffffffff823a6605 in dmu_brt_clone <br> > (os=os@entry=0xfffff800c5ce4000, object=<optimized out>, <br> > offset=offset@entry=0, length=length@entry=207477, <br> > tx=tx@entry=0xfffff8071a108d00, bps=bps@entry=0xfffffe01e218c000, <br> > nbps=2, replay=0)<br> > at /usr/src/sys/contrib/openzfs/module/zfs/dmu.c:2303<br> > #6 0xffffffff8250f67c in zfs_clone_range (inzp=0xfffff804416ac000, <br> > inoffp=0xfffff800b81cb048, outzp=0xfffff806f58f03a0, <br> > outoffp=0xfffff800b8063048, lenp=lenp@entry=0xfffffe022f2bdbf0, <br> > cr=0xfffff8000a6fe600)<br> > at /usr/src/sys/contrib/openzfs/module/zfs/zfs_vnops.c:1326<br> > #7 0xffffffff8234b3ba in zfs_freebsd_copy_file_range <br> > (ap=0xfffffe022f2bdc48) at <br> > /usr/src/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c:6294<br> > #8 0xffffffff8079f443 in VOP_COPY_FILE_RANGE <br> > (invp=0xfffff804416cb1c0, inoffp=0xfffff800b81cb048, <br> > outvp=0xfffff806f51d3380, outoffp=0xfffff800b8063048, <br> > lenp=0xfffffe022f2bdd48, incred=0xfffff8000a6fe600, flags=<optimized <br> > out>,<br> > outcred=<optimized out>, fsizetd=<optimized out>) at <br> > ./vnode_if.h:2385<br> > #9 vn_copy_file_range (invp=invp@entry=0xfffff804416cb1c0, <br> > inoffp=inoffp@entry=0xfffff800b81cb048, <br> > outvp=outvp@entry=0xfffff806f51d3380, <br> > outoffp=outoffp@entry=0xfffff800b8063048, <br> > lenp=lenp@entry=0xfffffe022f2bdd48, flags=flags@entry=0,<br> > incred=0xfffff8000a6fe600, outcred=0xfffff8000a6fe600, <br> > fsize_td=0xfffffe022925b3a0) at /usr/src/sys/kern/vfs_vnops.c:3087<br> > #10 0xffffffff8079a070 in kern_copy_file_range <br> > (td=td@entry=0xfffffe022925b3a0, infd=<optimized out>, <br> > inoffp=0xfffff800b81cb048, inoffp@entry=0x0, outfd=<optimized out>, <br> > outoffp=0xfffff800b8063048, outoffp@entry=0x0, len=9223372036854775807,<br> > flags=0) at /usr/src/sys/kern/vfs_syscalls.c:4973<br> > #11 0xffffffff8079a178 in sys_copy_file_range (td=0xfffffe022925b3a0, <br> > uap=0xfffffe022925b7a0) at /usr/src/sys/kern/vfs_syscalls.c:5011<br> > #12 0xffffffff80a97aa3 in syscallenter (td=0xfffffe022925b3a0) at <br> > /usr/src/sys/amd64/amd64/../../kern/subr_syscall.c:188<br> > #13 amd64_syscall (td=0xfffffe022925b3a0, traced=0) at <br> > /usr/src/sys/amd64/amd64/trap.c:1194<br> > #14 <signal handler called><br> > #15 0x000007fbb2da4ada in ?? ()<br> ><br> ><br> > and disabling bclone does appear to allow me to finish buildworld / <br> > buildkernel.<br> ><br> > The pool didn't have redaction_list_spill enabled.<br> ><br> > The ASSERT3U(nbps, ==, numbufs); in dmu_brt_clone was added when block <br> > clone is first implemented.<br> ><br> > It seems that I am the only person who is seeing this as of today. It <br> > seems that block clone was indeed being used for some data:<br> ><br> > saturn bcloneused 1.18M -<br> > saturn bclonesaved 1.21M -<br> > saturn bcloneratio 2.02x -<br> ><br> > The pool have dedup enabled for some datasets.<br> ><br> > Any suggestions? (In extreme cases I can recreate the storage pool <br> > from backup or copy the data somewhere else, then recreate the pool, <br> > then copy data back, but I'd like to avoid that if possible)<br> ><br> > Cheers,<br> ><br> ><br> >><br> >> mm<br> >><br> >> On 04/11/2023 13:35, Mark Millard wrote:<br> >>> On Nov 4, 2023, at 04:38, Mike Karels <<a href="mailto:mike@karels.net" target="_blank" moz-do-not-send="true" class="moz-txt-link-freetext">mike@karels.net</a>> wrote:<br> >>><br> >>>> On 4 Nov 2023, at 4:01, Ronald Klop wrote:<br> >>>><br> >>>>> On 11/4/23 02:39, Mark Millard wrote:<br> >>>>>> It looks to me like releng/14.0 (as of 14.0-RC4) still has:<br> >>>>>><br> >>>>>> int zfs_bclone_enabled;<br> >>>>>> SYSCTL_INT(_vfs_zfs, OID_AUTO, bclone_enabled, CTLFLAG_RWTUN,<br> >>>>>> &zfs_bclone_enabled, 0, "Enable block cloning");<br> >>>>>><br> >>>>>> leaving block cloning effectively disabled by default, no<br> >>>>>> matter what the pool has enabled.<br> >>>>>><br> >>>>>> <a href="https://www.freebsd.org/releases/14.0R/relnotes/" rel="noreferrer" target="_blank" moz-do-not-send="true" class="moz-txt-link-freetext">https://www.freebsd.org/releases/14.0R/relnotes/</a>; also reports:<br> >>>>>><br> >>>>>> QUOTE<br> >>>>>> OpenZFS has been upgraded to version 2.2. New features include:<br> >>>>>> •<br> >>>>>> block cloning, which allows shallow copies of blocks in file <br> >>>>>> copies. This is optional, and disabled by default; it can be <br> >>>>>> enabled with sysctl vfs.zfs.bclone_enabled=1.<br> >>>>>> END QUOTE<br> >>>>>><br> >>>>><br> >>>>> I think this answers your question in the subject.<br> >>>> I think so too (and I wrote that text).<br> >>> Thanks for the confirmation of the final intent.<br> >>><br> >>> I believe this makes:<br> >>><br> >>> QUOTE<br> >>> author Brian Behlendorf <<a href="mailto:behlendorf1@llnl.gov" target="_blank" moz-do-not-send="true" class="moz-txt-link-freetext">behlendorf1@llnl.gov</a>> 2023-05-25 20:53:08 <br> >>> +0000<br> >>> committer GitHub <<a href="mailto:noreply@github.com" target="_blank" moz-do-not-send="true" class="moz-txt-link-freetext">noreply@github.com</a>> 2023-05-25 20:53:08 +0000<br> >>> commit 91a2325c4a0fbe01d0bf212e44fa9d85017837ce (patch)<br> >>> tree dd01dfce6aeef357ade1775acf18aade535c6271<br> >>> . . .<br> >>> Update compatibility.d files<br> >>><br> >>> Add an openzfs-2.2 compatibility file for the next release. Edon-R <br> >>> support has been enabled for FreeBSD removing the need for different <br> >>> FreeBSD and Linux files. Symlinks for the -linux and -freebsd names <br> >>> are created for any scripts expecting that convention. Additionally, <br> >>> a symlink for ubunutu-22.04 was added. Signed-off-by: Brian <br> >>> Behlendorf <<a href="mailto:behlendorf1@llnl.gov" target="_blank" moz-do-not-send="true" class="moz-txt-link-freetext">behlendorf1@llnl.gov</a>> Closes #14833<br> >>> END QUOTE<br> >>><br> >>> technically incorrect in that compatibility.d/openzfs-2.2-freebsd<br> >>> should be distinct in content from compatibility.d/openzfs-2.2 so<br> >>> that block cloning would not be enabled.<br> >>><br> >>><br> >>>>>> Just curiousity on my part about the default completeness of<br> >>>>>> openzfs-2.2 support, not an objection either way.<br> >>>>>><br> >>>>><br> >>>>> I haven't seen new issues with block cloning in the last few weeks <br> >>>>> mentioned on the mailing lists. All known issues are fixed AFAIK.<br> >>>>> But I can imagine that the risk+effect ratio of data corruption is <br> >>>>> seen as a bit too high for a 14.0 release for this particular <br> >>>>> feature. That does not diminish the rest of the completeness of <br> >>>>> openzfs-2.2.<br> >>>>><br> >>>>> NB: I'm not involved in developing openzfs or the decision making <br> >>>>> in the release. Just repeating what I read on the lists.<br> >>>> There was another block cloning fix in 14.0-RC4; see the commit log.<br> >>>> Maybe there will be no more issues, but it seems that corner cases <br> >>>> were<br> >>>> still being found recently.<br> >>> Looks like I'll stay at openzfs-2.1 pool features until there is<br> >>> a release that no longer has the default status:<br> >>><br> >>> 0 for sysctl vfs.zfs.bclone_enabled<br> >>><br> >>> I use main [so: 15 now] but only enable openzfs-2.* pool features<br> >>> supported by default on some FreeBSD release, that has an accurate<br> >>> compatibility.d/openzfs-2.*-freebsd file.<br> >>><br> >>> ===<br> >>> Mark Millard<br> >>> marklmi at <a href="http://yahoo.com" rel="noreferrer" target="_blank" moz-do-not-send="true">yahoo.com</a><br> >>><br> >>><br> >><br> ><br> <br> </blockquote> </div> </div> </blockquote> </body> </html> --------------hcDCO5UsWNX1hkTPj6uApmeT--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?77ad9593-34a8-48dc-8533-aafb852f1d19>