How performant would #ZFS be on a #Thinkcentre with an i3 -4160 and 4 GiB of DDR3 RAM?
Would it be usable as a headless remote storage/backup box with 3x 1TB SATA SSDs in RAIDZ1?

If I can get #FreeBSD to boot on it that is. #homelab #selfhosted #selfhosting

(boosts welcome)

1/6
I’ve been thinking through disaster recovery with #NixOS #flakes, impermanence, and #ZFS.
A flake defines a system declaratively, but assumes storage (e.g. ZFS pools) and persisted data (e.g. /persist) already exist. Flakes don’t create storageβ€”they only configure it. Right?
GitHub - acidvegas/zpulse: Real-time ZFS & SMART monitoring dashboard for multi-server home racks

Real-time ZFS & SMART monitoring dashboard for multi-server home racks - acidvegas/zpulse

GitHub

I tend to keep too many #OpenZFS #ZFS boot environments. Case in point, my DNS server:

BE Active Mountpoint Space Created
15s-2026-01-21_01 NR / 30.4G 2026-01-21 12:53
default - - 735M 2020-03-22 05:06
master-2020-03-22_01 - - 948M 2020-03-22 09:23
master-2020-06-19_01 - - 1001M 2020-06-19 03:00
master-2020-10-20_01 - - 1017M 2020-10-20 09:11
master-2021-01-11_01 - - 1014M 2021-01-11 12:21
master-2021-04-05_01 - - 95.4M 2021-04-05 19:43
master-2021-06-25_01 - - 1.20G 2021-06-25 10:58
master-2021-08-31_01 - - 1.26G 2021-08-31 08:13
master-2021-10-07_01 - - 1.36G 2021-10-07 10:39
master-2022-03-25_01 - - 1.25G 2022-03-25 18:14
master-2023-01-22_01 - - 1.41G 2023-01-22 06:15
master-2023-06-02_01 - - 1.52G 2023-06-02 20:11
master-2023-08-27_01 - - 1.51G 2023-08-27 12:34
master-2023-11-09_01 - - 1.55G 2023-11-09 14:08
master-2024-03-29_01 - - 1.60G 2024-03-29 18:47
master-2024-10-01_01 - - 1.54G 2024-10-01 16:29
master-2024-11-30_01 - - 1.46G 2024-11-30 17:58
master-2025-08-06_01 - - 2.05G 2025-08-06 14:00
master-2025-12-27_01 - - 1.40G 2025-12-27 18:09

Still have the original boot environment created six years ago.

I'm back with another #FreeBSD question, specifically about `periodic`. I'd like to use `periodic daily` to occasionally scrub my ZFS pools by setting `daily_scrub_zfs_enable` to `"YES"` in `/etc/periodic.conf`. This works great on a server, because it's always on, but not so great on my laptop, because the cron is scheduled at night. I could change the time, but then it'll only run if my laptop is on at the right time.

Is there a way to schedule this so the `periodic` crons are run after booting, in case there are "missed" runs? What is the recommended approach here, use e.g. Xfce's autostart applications for this?

Edit: solved! The answer is anacron

Thanks for reading :)

#cron #periodic #zfs

My day is rapidly going from bad to worse. Not having fun right now as the spare disk in the cupboard is not the right size. Luckily it is larger than the faulty one so I should be able to mess around with gpart to get things going.
There is only nightly backups and my mac time machine backups on the pool so I have a bit of time
#ZFS #FreeBSD #sysadmin #diskfailure

20 hours for "zpool attach" a single drive to #raidz1 in my case =) Turns out you can't "attach" two drives at once, so there will be two resilvering processes after each drive attachment. 40 hours in total.

#zfs #homelab

Striped poll consists of two drives as well. At the end, I should have #Raidz1 poll with 4 drives. Finally, having the same size.

#homelab #zfs #zpool #nas

"ZFS with a striped configuration can be more space-efficient compared to other setups like RAIDZ, as it does not require additional space for parity. However, the actual usable space will depend on the number of disks and the specific configuration used."

With this in mind, I started to migrate data from striped #zpool to #Raidz1

There was around 10TB of data to move. But after moving all that data, it took around 20TB on a two-drive raidz1. Some #BlackMagicFuckery

#zfs #nas #homelab

So after expanding my #zfs mediapool yesterday, today I will implement a new backup scheme (hopefully). I will use a 6TB USB3 disk for automated backups of my most important data.