#zfs

helps to remember MAKEDEV when you're tinkering with ZFS pools #NetBSD #ZFS

I actually read the #FreeBSD Handbook's chapter on how FreeBSD bootstraps. It's stuck in 2010. I'm not kidding. There's stuff in there that changed in the FreeBSD 7.3 release notes. FreeBSD can boot on UEFI firmware nowadays, and directly from ZFS datasets. There's a gpart bootstrap command, and syscons does not use bright white for kernel messages any more. In the meantime, the Handbook is telling people how to accidentally destroy their MBR or GUID partition tables with bsdlabel. Do you think that we can get the Handbook people to catch up with the actual operating system, that is 16 years ahead? (-: https://github.com/jdebp/freebsd-doc/blob/main/documentation/content/en/books/handbook/boot/_index.adoc #UEFI #ZFS

#bhyve with #ZFS rollback makes testing boot environments for #openzfsmastery sooo much easier. Snapshots on the hypervisor. Snapshots in the VM. Snapshots EVERYWHERE. #sysadmin

It's probably time to have a deeper look at #Zelta #ZFS #FreeBSD

@anparker@mastodon.bsd.cafe @mroach@ublog.mroach.com @subnetspider@mastodon.bsd.cafe Unused RAM is wasted $$$$ #ZFS The consumption of RAM will back off if it is needed for other purposes but why not have it improve storage performance if the RAM is just running idle in a host.

One of my NVMe SSDs dropped out of several arrays, again. 👉 a Transcend MTE220S 256er As an IT guy myself, i've just rebooted the box. That made matters worse, because the dodgy SSD didn't enumerate (/dev/nvme*), and i was left with 3/4 SSDs, incorrectly enumerated. My #ZFS pools re-imported fine, albeit in a DEGRADED state. zpool replace is bullshit! 🤬 ..it NEVER wörks! I had to power-cycle the box, and zpool detach/remove, then add & resilver shit! SMART reports look fine! 🧵1/n..

Well I gotta say, moving #MastodonAU 's database to a pure SSD storage backend has made a massive difference. Whooohoo, she's much faster! The storage servers are a pair of #Dell R640s, with 2x Gold 6150 (2.7GHz 18 Cores each) and 256GB of RAM. Each one has 8 x 1TB SSDs, in a #ZFS raidz2 pool, and they automatically replicate between themselves. I am doing my best to avoid any more visits from the #fuckup-fairy Now to finish #MastodonNZ

With Apple phasing out AFP in macOS 27 and Time Capsules officially reaching end-of-life, it's time to move network backups to proper SMB. If you run a FreeBSD server, you can build a fast, rock-solid, and secure Time Machine target powered by ZFS and Samba - neatly isolated inside a FreeBSD jail using Bastille. https://it-notes.dragas.net/2026/01/28/time-machine-freebsd-jail/ Keep your macOS backups running smoothly via SMBv3 (with full vfs_fruit support) and full dataset quota control on ZFS! #FreeBSD #macOS #TimeMachine #ZFS #BastilleBSD #Samba #SysAdmin #Backup #OwnYourData #SelfHosted #BSD #RunBSD #OwnYourData

How much disk space does it take to reliably track & scan all the important open-source software releases published throughout the year? About 120TB. Thanks to #FreeBSD, #ZFS, #WesternDigital, and #Ampere for being awesome. :)

Nearly two years ago I noticed one NVME disk had endurance gone down to zero. It has been part of ZFS RaidZ3 six drive pool under constant write load. Normally I would replace such drive. But this time I decided to see how long it takes to break it. And the answer is - almost two years. The disk receives few MB/s writes *minimum* with hundreds MB/s peaks, maybe 100TB writes per year. I'll try to dig some numbers later. So, looks like there is ample headroom in NVME write endurance claims. (statistics from sample of one 🤣 ) I think I'll let the pool run sans one drive.It should survive two more drive failures. And there are daily/weekly/monthly backups available. Maybe buy new after AI bubble bursts. #homelab #zfs #diskfailure

If Scrubs Hurt, Your ZFS Design Is Broken https://klarasystems.com/articles/if-scrubs-hurt-your-zfs-design-is-broken/ #HackerNews #Tech #ZFS

Latest 𝗩𝗮𝗹𝘂𝗮𝗯𝗹𝗲 𝗡𝗲𝘄𝘀 - 𝟮𝟬𝟮𝟲/𝟬𝟳/𝟮𝟳 (Valuable News - 2026/07/27) available. https://vermaden.wordpress.com/2026/07/27/valuable-news-2026-07-27/ Past releases: https://vermaden.wordpress.com/news/ #verblog #vernews #news #bsd #freebsd #openbsd #netbsd #linux #unix #zfs #opnsense #ghostbsd #solaris #vermadenday

All #BSD and #ZFS wizards alert Please focus all your mojo on helping the people who posted the ZFS/BSD admin job that I just applied for to see that I am their dream come true.

Latest 𝗩𝗮𝗹𝘂𝗮𝗯𝗹𝗲 𝗡𝗲𝘄𝘀 - 𝟮𝟬𝟮𝟲/𝟬𝟳/𝟮𝟬 (Valuable News - 2026/07/20) available. https://vermaden.wordpress.com/2026/07/20/valuable-news-2026-07-20/ Past releases: https://vermaden.wordpress.com/news/ #verblog #vernews #news #bsd #freebsd #openbsd #netbsd #linux #unix #zfs #opnsense #ghostbsd #solaris #vermadenday

A couple of days ago, few hours before leaving home for a four days trip I decide to upgrade one of my home servers to FreeBSD 15.1. Nothing difficult. After upgrade the machine start rebooting in a loop. I pick up my serial cable and see the boot process crash when mounting ZFS panic: zfs(panic): zfs: rt={spa=zroot vdev_guid=546977716602668702 ms_id=9 ms_freed}: adding segment (offset=2431c0000 size=1000) overlapping with existing one (offset=2431c0000 size=1000) 😬 that's terrible. I relay on this machine to access my nextcloud where all my trip tickets are stored. Hopefully ZFS saves my day! I was able to boot back to the previous snapshot, and smoothly restore the service, but I probably need to start thinking in a replacement for this machine.... #FreeBSD #ZFS

@oxy@social.bsdlab.au @jpm@aus.social Solaris and #ZFS would be on my shopping list. I don't know what I'd do with UltraSPARC

Hey #FreeBSD Questions about migrating some dockerized services from Linux. *By the way, for those who noticed my #ZFS post from a day or two ago, good news: It wasn't the drive that was dead! It was either motherboard, of which I have functional spares at hand, or the 750w power supply, but I strongly suspect the motherboard, which is old enough to have classic PCI slots on it. The array is absolutely fine and the server's back up with my newest mobo and sadly, an i3. If you have an LGA1151 i7 CPU (either 6th-7th or 7th-8th gen Intel, can't recall) that you're not using and want to donate it to the chezpants.ca cause, dm me * Anyways, I decided to give FreeBSD another go, since this is primarily a ZFS-based file server, and since I started listening to @25admins@mastodon.social they've been seducing me back to the ummm Dark Side. My first and only Sysadmin job was BSD-heavy, and it truly is a vastly superior server experience than anything else out there. But I was running several services on that metal with Docker as well, and while in theory I can install Podman and bring those back up, I would rather embrace native solutions where possible. The Jellyfin container is already running the pkg version, flawlessly. I got a little confused when attempting to setup ports, so I'm just going with pkg for now. Other stuff to replace is FreshRSS, Pi-Hole, Mealie, and I intend to get some kind of ebook utility with a web UI up. Other stuff too, but those are the daily use ones that I need to replace like today. I have heard that Blocky is the predominant Pi-Hole alternative? I haven't looked into Mealie yet, and FreshRSS doesn't have a supported BSD install, so I'd rather try out some other RSS util that does, if there's one that's worth a damn. WebUI as well, please - I am aware there's a terminal based one and I might install it for shits and giggles but I want to see the pictures and whatnot, generally. Oh, and also I'm gonna go ahead and tag @stefano@mastodon.bsd.cafe cause he seems to be a pretty good authority on running services on BSD - in fact I think bsdcafe has FreshRSS, even?

How to Build a Minimal ZFS NAS Without Synology, QNAP, TrueNAS https://neil.computer/notes/how-to-setup-minimal-zfs-nas-without-truenas/ #HackerNews #Tech #ZFS

Thought I was real clever, building a 4x16TB #zfs array for our server, umm... guess it was a good six years ago now. They cost CDN$400 round then. Think of the hydro, I said! Today, I'm pulling the old 6x4TB that that array replaced, pruning what data I can from the currently-degraded 16TB array in order to transfer it over, and then rebuilding it as a 3-drive zpool for non-critical data, cause holy shit we cannot afford 16TB drives right now. I might even just use em as straight 16TB drives, why even bother with the pretense that I could replace one... (I know that the math adds up the same when replacing hard drives, TB for TB, but we can handle the failure of a 4TB drive as an unanticipated money hit. A 16TB drive, that's been elevated to "tax time luxury purchase" costs)