I use Claude Code via Ollama to manipulate files and folders on my MacBook.

I’ve tried smaller models like Gemma 4 and Qwen 2.5 Coder in 7B, but they don’t work well (or maybe I just don’t know how to use them properly).

I’ve also tried larger 14B models, such as Qwen2.5‑Code‑14B, but when I run a prompt, my MacBook slows down a lot, sometimes freezes for a few seconds, and I have to wait several minutes. I was wondering if this is normal.

Hello,

I'm looking for a 3-in-1 charging station for my Apple products (iPhone, Apple Watch, and AirPods). I've been checking out quite a few comparisons and these 3 products have caught my attention, but I don't know which one to choose.

On one hand, there's the ESR 3-in-1 Travel which seems lightweight and has similar features to the standard station (according to Cheval Lim's video).

On the other hand, there's the Anker model. I haven't found any concrete tests regarding heating, charging speed, etc.

I wanted to ask if anyone has feedback on these 3 products and which one you'd recommend?

A few clarifications: I don't travel much, but if I do go somewhere (to someone's place), I'll take this charger with me (I have a power bank for transport trips or other situations...). I'm also looking for (if possible) a station that can charge slowly when charging at night (via a button or something similar).

Hello, I recently installed TrueNAS and I’m in the process of setting up all my services again, like Radarr, Prowlarr, qBittorrent, etc.

I wanted to know whether it would be better to use Dockge for this or to install the applications provided directly by TrueNAS. I’d also like to route the torrent traffic from my qBittorrent client through a VPN (I currently have a Proton VPN subscription).

Previously, I had a Proxmox server with a Linux VM running Dockge, which managed all my containers. I recently realized that what I had set up wasn’t ideal, so I moved to TrueNAS while keeping the pools I already had.

I currently have two pools where I store media (movies and TV shows). I created a second pool because the first one was full.

Recently, I did a big cleanup, so I now have free space on both pools. I’d like to download torrents to both pools, switching to the other one once the first becomes full.

Background (please be kind, I'm new to this)

I'm a complete beginner in the server/homelab world. A few months ago I decided to set up a home server to store my media collection and share it with my family through Plex. I built the whole thing with heavy help from AI assistants (ChatGPT, Claude) (I know this is probably not what the community recommends, and I understand why now. But I really wanted to set up my own server and this was the only way I saw to get started.) I don't have deep Linux/ZFS knowledge.

My setup has been running for several months without major issues, until few days ago. I wanted to watch a movie on Plex but it was extremely slow, the movie wouldn't start, the interface was laggy. I went to check my Proxmox dashboard and noticed TrueNAS was in a weird state. I tried to reboot it, and that's when everything went downhill.

I then spent multiple hours with AI assistance trying to fix it, and I'm now pretty sure I've made things much worse. I'm here because I think I need real humans with real ZFS expertise.

Hardware setup

• Proxmox VE 8.4.14 on a single physical box
• CPU: 4 cores allocated to TrueNAS
• RAM: 16GB total on host (single slot, can't easily upgrade), 10GB to TrueNAS VM
• Boot/system: 464GB NVMe with LVM thin (very full, PFree = 0)
• TrueNAS SCALE 24.10.2 as a QEMU VM with disks passed through
• 1 VM Linux with my containers (Dockge with qBittorrent, Sonarr, Radarr, etc.) (3GB RAM)
• 1 LXC Container (Plex) (2GB RAM)

I have 2 Media pools:

• 3 x 12TO (the one I broke) "Media 12TO"
• 4 x 4TO "Serveur_Wilfred"

Before this incident, my setup had known pain points:

• 14 CKSUM errors on Media 12TO a few weeks ago, one corrupted file which I deleted and ran zpool clear
• middlewared had crashed multiple times in recent weeks due to RAM pressure (ARC eating 7+ GB on 8GB allocation, OOM killing middlewared, I bumped TrueNAS to 10GB after that)

What happened

Day 1, the unclean shutdown

TrueNAS lost power or froze sometime during the night (the last valid uberblock timestamp on disk confirms a shutdown around that time). I don't know the exact cause, maybe a power blip, maybe a crash, I just noticed it the next morning when Plex was broken.

Day 2, my attempts to fix it

At boot, TrueNAS hung:

• ix-zfs.service stayed stuck for 15+ minutes, then failed
• Media 12TO import got stuck on "Syncing ZIL claims" phase (confirmed in /proc/spl/kstat/zfs/dbgmsg)
• zpool import process ended up in D state (uninterruptible sleep), unkillable
• spa_deadman warnings growing: "slow spa_sync: started 606 seconds ago to 2204 seconds" alternating between the 3 disks
• No kernel I/O errors on the disks (dmesg completely clean on sdg/sdh/sdi)
• Serveur_Wilfred imports successfully every time, no issues

Multiple reboots did not help, same hang every time on Media 12TO.

The commands I tried (and where I think I broke things)

All commands below were issued over the course of the day, with the VM rebooted into init=/bin/bash via GRUB between tries to avoid the auto-import hang.

1. zpool import -N "Media 12TO" gave hung in D state
2. zpool import -F -N "Media 12TO" gave hung
3. zpool import -FX -N "Media 12TO" gave hung. I think this is where I destroyed things.
4. zpool import -o readonly=on -N "Media 12TO" succeeded, but zpool list shows 0 ALLOC on a 32.7T pool
5. zpool import -o readonly=on -T <txg> -f "Media 12TO" hung >15min
6. zpool import -o readonly=on -T <older_txg> -f "Media 12TO" hung >15min
7. zpool import -o readonly=on -o cachefile=none -fFX "Media 12TO" imports, still 0 ALLOC

I also restored a vzdump backup of the TrueNAS VM (system disk only) as a new VM while keeping the original stopped, reattached the passthrough disks to the new VM and tried imports from there. Same results.

Current state

• VMs are both stopped cleanly
• All 7 data disks physically healthy, no SMART errors, no I/O errors, all ONLINE in zpool import output
• Uberblocks intact on disk
• Pool imports in readonly mode, but reports 0 ALLOC and 0% CAP (when it had 19TB of data 24h ago)

My questions for you

1. Is Media 12TO truly destroyed, or are my 19TB of data still physically on disk but just unreachable because -FX trashed the metadata pointers?
2. Is there a zdb -e technique to inspect datasets/MOS without importing the pool, to confirm whether data blocks are still out there?
3. Would echo 1 > /sys/module/zfs/parameters/zfs_recover before an import attempt help, or is it too late at this point?
4. Is zpool import -T with a TXG from before my -FX (the earliest available uberblock) worth trying again, or is that just repeating what I already tried?
5. Given the disks are physically fine and this is purely metadata damage, what's the realistic path forward?
   • Is there any chance of DIY recovery with more advanced zdb commands I haven't tried?
   • Or is this a professional recovery job at this point ?

What I'm asking from you

I know I caused this myself by running -FX without readonly=on based on AI suggestion without understanding it. I'm not looking for blame, I'm looking for any path forward before I accept the loss. If the answer is "your data is gone, recreate the pool", I'll accept it, but I want to hear it from people who actually know ZFS internals.