Date: Wed, 2 Oct 2019 02:04:40 +0000 From: Shiva Bhanujan <Shiva.Bhanujan@Quorum.com> To: Eric Borisch <eborisch@gmail.com> Cc: Fabian Keil <freebsd-listen@fabiankeil.de>, "freebsd-fs@freebsd.org" <freebsd-fs@freebsd.org>, Shiva Bhanujan <Shiva.Bhanujan@Quorum.com> Subject: RE: [ZFS] Potential silent data corruption when receiving incremental stream with block size change on receiving side Message-ID: <3A5A10BE32AC9E45B4A22F89FC90EC0701C3767B84@QLEXC01.Quorum.local> In-Reply-To: <CAMsT2=nO7e=E-1gFdA0bhiui2vt8_kMbjH2ghy6g4q8a_tu3LQ@mail.gmail.com> References: <20170807165118.72bde0a3@fabiankeil.de> <3A5A10BE32AC9E45B4A22F89FC90EC0701C3767B55@QLEXC01.Quorum.local>, <CAMsT2=nO7e=E-1gFdA0bhiui2vt8_kMbjH2ghy6g4q8a_tu3LQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Eric, thanks for the pointer. I tried the patch and it works. regards, Shiva ________________________________ From: Eric Borisch =5Beborisch=40gmail.com=5D Sent: Tuesday, October 01, 2019 5:03 PM To: Shiva Bhanujan Cc: Fabian Keil; freebsd-fs=40freebsd.org Subject: Re: =5BZFS=5D Potential silent data corruption when receiving = incremental stream with block size change on receiving side That sounds like a potential hole birth issue. You might consider upgrading to at least 10.4 or 11.1 and re-trying the = send; these added the vfs.zfs.send_holes_without_birth_time (default = true/1) tunable to avoid =22numerous hole_birth bugs=22. See this =5B1=5D = commit. - Eric =5B1=5D = https://svnweb.freebsd.org/base/releng/10.4/sys/cddl/contrib/opensolaris/ut= s/common/fs/zfs/dmu_traverse.c?revision=3D308086&view=3Dmarkup On Tue, Oct 1, 2019 at 5:51 PM Shiva Bhanujan = <Shiva.Bhanujan=40quorum.com<mailto:Shiva.Bhanujan=40quorum.com>> wrote: Hello, This seems to be an older email, but I seem to be hitting this in FreeBSD = 10.3. Following is the precise version - FreeBSD R530-Filer-20-194.quorum.net<http://R530-Filer-20-194.quorum.net>; = 10.3-RELEASE-p29 FreeBSD 10.3-RELEASE-p29 =2312 r335461M: Wed Jun 20 = 16:11:04 PDT 2018 = root=40R520-31-232.quorum.net:/usr/obj/usr/src/sys/QUORUM amd64 When I send an incremental snapshot, the md5sum of a couple of the files = in the source and the destination don't match. The previous incrementals = don't have a mismatch. Can somebody please advise on what I should be looking for? I've checked = the URLs mentioned here. The addition or removal of the -L option doesn't = have any change. Regards, Shiva ________________________________ From: = owner-freebsd-fs=40freebsd.org<mailto:owner-freebsd-fs=40freebsd.org> = =5Bowner-freebsd-fs=40freebsd.org<mailto:owner-freebsd-fs=40freebsd.org>=5D= on behalf of Fabian Keil = =5Bfreebsd-listen=40fabiankeil.de<mailto:freebsd-listen=40fabiankeil.de>=5D Sent: Monday, August 07, 2017 8:10 AM To: freebsd-fs=40freebsd.org<mailto:freebsd-fs=40freebsd.org> Subject: =5BZFS=5D Potential silent data corruption when receiving = incremental stream with block size change on receiving side The original message is attached because it is signed. ________________________________ _______________________________________________ freebsd-fs=40freebsd.org<mailto:freebsd-fs=40freebsd.org> mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-fs To unsubscribe, send any mail to = =22freebsd-fs-unsubscribe=40freebsd.org<mailto:freebsd-fs-unsubscribe=40fre= ebsd.org>=22 ________________________________
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3A5A10BE32AC9E45B4A22F89FC90EC0701C3767B84>