Finding old zfs pool on disks

Eight or nine years ago, I upgraded my main machine--in one of those upgrades that left the power cord all that was left, and maybe the keyboard (probably not; I'd been using a Model M).

The machine was meant for database crunching and development, but that never happened, and it pretty much sat unused.

A year or two later, I probably overwrote the mar with an ubuntu disk--the machine got hauled back out as part of an exchange involving restoration of an early 90s mac backup and reading those disks, and bouncing the results onto a modern mac.

Anyway, this once would have been mighty quad core machine with a gaming case, motherboard, and fancy graphics cards had four 1.5T 7200 rpm SATA drives.

The plan was a boot drive with FreeBSD for work, a stray linux du jour for odds and ends that still came up, probably a windows install (there's an ATA hard drive mounted in the case from a prior machine). and a 3tb zfs /home directory spread across 3 disks as raid.

Of course, what I intended, and what I did, might be two different things.

The loose SATA drive seems to have an ubuntu boot loader that doesn't know about freebsd (I tried to leave spaces between the drive for airflow, and I think I meant to mount it later, or replace the ATA, or some such).

I have no reason to believe that the SATA drives are connected to the same controller lines they used to be (or even if that would make a difference; it's been a while).

How can I find my old zfs /home? There are probably some things restored from some long-lost backups on it that I'd like to recover.

It looks like there is some version of "zfs import" that should tell me what's available, but I'm not seeing what that is from my googling and the manages--and I'm worried that one mistyped command could shoot my feet off.

I believe that this is the first release of FreeBSD in which zfs was not experimental; it would have been sometime in 08 or 09.

hawk
 
If you attach disks that contain a ZFS pool, or were part of a pool, to a new computer, zpool import should scan all disks and show you what it can find.

If enough disks are present for it to actually read data from the pool, you can run zpool import [-N] {name} to import the pool into the system. You may want to use the -N argument to not automatically mount any filesystems - If you have a /home mount in the pool for example, you probably don't want that mounting over the top of your current /home directory.

Once a pool is imported you can use mount -t zfs pool/some/dataset /some/path to mount a dataset somewhere in order to recover data from it.
 
That did it, thank you.

It does not that I have an archaic pool that will have to be updated :)

Now that I've found the dhclient command (which I new in another century), I've connected, done a mount_union set, which let me install rsync, and I appear to be in business.

Thank you very much

hawk
 
Back
Top