Date: Thu, 23 Oct 2014 11:35:11 +0200 From: Kazik CC <dmilith@gmail.com> To: freebsd-current@freebsd.org Subject: Re: ZFS: i/o error - all block copies unavailable #2 Message-ID: <9711EF60-74E0-46E4-A89E-CD0486F3C457@gmail.com> In-Reply-To: <20141022153846.GA7092@brick.home> References: <B6E7646B-9A7D-43BF-B25A-5488D37B5CA5@gmail.com> <20141022153846.GA7092@brick.home>
next in thread | previous in thread | raw e-mail | index | archive | help
In old post about same issue on 8.x tree, from 2009, I saw the patch = that fixed some integer overflows for large disk sizes/ datasets for = raidz.. maybe it=E2=80=99s something similar here, but for huge stripes? Also, if it might be =E2=80=9Cbootloader launching too fast, before BIOS = actually see all the drives=E2=80=9D=E2=80=A6 so, maybe there=E2=80=99s = a way, to slow bootloader down in case of huge drive? 65TiB is pretty huge virtual drive. Besides that, controller launches all 36 drives properly before boot = process is launched.. The zpool is untouched, I scrubbed it without any = issues. I=E2=80=99m pretty new to fBSD codebase, where can I look into the code = for answers? > On 22 Oct 2014, at 17:38, Edward Tomasz Napiera=C5=82a = <trasz@FreeBSD.org> wrote: >=20 > On 1022T1633, Kazik Chujwielki wrote: >> =E2=80=A6 >>=20 >> So, I tried with upgrading to 9.3, installing bootcode from 9.3, then = I booted 10.0, installed boot code again.. nothing helps. >>=20 >> Still ends on = http://s.verknowsys.com/1a4f147cbb215da9b16eb61a7e9de9ed26470338.png = <http://s.verknowsys.com/1a4f147cbb215da9b16eb61a7e9de9ed26470338.png>; >>=20 >> I also tried scrubbing my pool, from ZFS point of view everything is = just "fine=E2=80=9D. No data errors.. it=E2=80=99s machine with ECC RAM, = so corruption shouldn=E2=80=99t take place.. >=20 > Does the scrubbing work under 10.0, the same kernel that fails to = boot? > How did you reinstall the bootcode? >=20
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9711EF60-74E0-46E4-A89E-CD0486F3C457>