Re: zfsutils migration (Re: new ZFS installs completely broken in Wheezy/Sid)
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello,
On 13.12.2011 18:51, Robert Millan wrote:
>> S1. upgrade to kfreebsd-8 8.2, and would need zfsutils 8.2
>
> kfreebsd 8.2 is usable with zfsutils 8.1, they've got a different ZFS
> version but AFAICT they use the same ABI (14 vs 15).
...
>> The only other option I see is to split zfsutils so there is a separate
>> version for installing+running kfreebsd-8 8.2, although that kernel
>> would disappear anyway when 8.3 replaces it.
>
> That's a big hassle, and it'd only be temporary.
Note, I didn't find any /real/ problem when using zfsutils 8.3 with a
8.2 kernel. I didn't try it with 8.1 but I don't see why it would be
different there.
Using newer utils does not break anything really, the only problem seems
to be that newer versions of zfsutils do not support "zfs volinit"
anymore, whereas the older kernel more or less waits for this call from
the user land to export volumes properly. Thus, when booting kfreebsd
with an older kernel but newer zfsutils some volumes might be missing on
boot. For the 8.3/9.0 branch this seems to happen automagically.
If we could find a workaround by emulating this "volinit" call when an
older kernel is detected, I guess we could happily use newer utils along
with older kernels. I haven't looked at the code, but that's probably
just some ioctl we could emulate easily.
- --
with kind regards,
Arno Töll
IRC: daemonkeeper on Freenode/OFTC
GnuPG Key-ID: 0x9D80F36D
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/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=tZjC
-----END PGP SIGNATURE-----
Reply to: