Re: How does one resize/expand a softraid RAID1?
Chris Ross <[email protected]> Wed, 29 Jul 2026 17:07:42 -0400
| Newsgroups | gmane.os.openbsd.misc |
|---|---|
| Message-ID | <[email protected]> |
> On 14 Jul 2026, at 05:21, Stuart Henderson <[email protected]> = wrote: >=20 > On 2026-07-13, Chris Ross <[email protected]> wrote: >> [a failed disk ; softraid mirror ; docs for bioctl ; OpenBSD] >>=20 >> Is there a way to replace a failed drive with a larger one, rebuild = the >> mirror onto it, then replace the smaller remaining drive with one of = equal >> size and rebuild onto that? [=E2=80=A6] >>=20 >=20 > Not supported with softraid. You can't create a "1-disk RAID1" and add > the second drive later either. Unless you want to get dirty with = poking > at low level data structures, you'll either need to have both new = drives > and the remaining working old drive connected at the same time and > dump/restore, or dump to external media + swap drives + restore, etc. That=E2=80=99s a shame. Would it be particularly difficult inside of = softraid to get to the equivalent of ZFS=E2=80=99 autoexpand behavior? It seems = like it shouldn=E2=80=99t be to tell the softraid device to just understand that = it=E2=80=99s the size of the smallest member of the mirror. At least for RAID1 it = shouldn=E2=80=99t be more complicated than that. Then, what the FS layer does underneath that is a separate issue, but mostly a solved problem for FFS already. I don=E2=80=99t have enough bays in this system to build a whole other = RAID1 and copy the data, so I=E2=80=99ll have to take it offline for a while = and move data around the network. In this case, I=E2=80=99m looking at less than = a TB so it won=E2=80=99t be a problem, but it would be if there were lots of = data. Thanks for the answers all. -Chris=