Summary of issues how I fixed them so far

[mattlangis]

So I had almost all the issues that were raised over the last few months after migrating from OMV to TrueNAS.

Sadly, TrueNAS doesn't support virtiofs, so I couldn't simply install Ubuntu and then start from scratch.
What @JonathanTreffler did is the best out there but I don't think it was expected to be used on so many different envrionments with different filesystems, ACL, etc ...

That being said, here my finding over the last couple of days:

1. The best image I found so far is the tag ubuntu22 one. It seems to get some traction lately, not sure if @JonathanTreffler is working on it, but thanks a bunch for that.

2. The /config mounting is the biggest issue I had. It seems it makes installing Wine problematic for some reason. It gave me the infamous blue screen almost all the time. I since then moved to an harder to back docker volume. Since I did that, I never had a single blue screen and I can repeat the installation process, always.

services:
  backblaze:
    image: tessypowder/backblaze-personal-wine:ubuntu22
    container_name: backblaze-personal
    ...
    volumes:
       - backblaze_data:/config
    ...

volumes:
  backblaze_data:

More details (you can skip this section if you don't care about the findings): It seems that when sharing a volume with the host, Wine and specially, dotnet for Wine doesn't install properly. It seems to make the whole container unstable. I'm by far not a Linux expert, but I think it's related to the way Wine do some symbolic links (or something similar) under the hood. Or it's around permissions, but I just can't figure it out due to my lack of knowledge in that area. What I know is that using volume gives the ability to still upgrade the container without losing settings. It seems to also cause the **Windows XP issue, because when winetricks install dotnet, they switch to WinXP. Then it crashes or something and never go back to Win7+

3. I do get the hanging backup issue too. I didn't find the cause yet, but it might be related to the way the encrypted data is put in RAM. I'm not yet 100%, but I did a script that is run with the health checks. What it does is if it find the infamous ca***\bz_done_*** stuck for a few minutes, it force kill the container.

Example of health check

healthcheck:
      test: ["CMD", "/bin/bash", "/usr/local/bin/backblaze_health_check.sh"]
      interval: 30s      
      timeout: 10s     
      retries: 1        
      start_period: 60s  
    restart: on-failure

If anyone is insterested on the script, I would be glad to share it.

4. Drives not showing up I think that @JonathanTreffler pretty much summarized it here, you need to do a docker exec --user app backblaze_personal_backup ln -s /backup_volume/ /config/wine/dosdevices/d: (or ln -s /backup_volume/ /config/wine/dosdevices/d: from within the container) and then restart the container

5. Permissions/ACL: I did not solve them yet. I went with what @focher proposed, to use user and group ID 0 (root:root) for now. But I think when using ZFS, it would be better to have a dedicated user:group for backups that have all the read permissions on all datasets.

6. Sub Datasets (children) in a Dataset (parent) seems to prevent BB from backuping the parent Dataset. That one I'm not sure if it's related to how datasets reports their files in the filesystem, but it might be related. When I have a dataset with 2 datasets and a random folder, it will only backup the random folder.

---

That's where I sit so far. If anyone has anything to add, let me know :)

A special thanks to @focher for all his comments, it helped me investigate further!

[JCStrainer]

Hello, woud you mind sharing your healthcheck script?

I followed your instructions and used ubuntu22 flag as well as a volume, unfortunately my setup still hangs on

[app ] + wine64 install_backblaze.exe

and the blue screen never disappears.

Thank you!

[mattlangis]

It's far from perfect, but it's a good start. It helped me backup a first portion of my backup

#!/bin/bash

# Backblaze Health Check Script
# Monitors for hanging "ca*/bz_done_*.bzff" transmissions
# Returns exit code 1 if stuck for more than 2 minutes

set -euo pipefail

# Configuration
OVERVIEWSTATUS_FILE="${BACKBLAZE_DATA_DIR:-/config/wine/drive_c/ProgramData/Backblaze/bzdata}/overviewstatus.xml"
STATE_FILE="/tmp/backblaze_hang_state"
HANG_THRESHOLD_SECONDS=120  # 2 minutes
LOG_PREFIX="[Backblaze Health Check]"

# Logging function
log() {
    echo "$LOG_PREFIX $(date '+%Y-%m-%d %H:%M:%S') - $1" >&2
}

# Function to extract XML attributes using grep/sed
get_xml_attr() {
    local file="$1"
    local attr="$2"
    
    if [[ ! -f "$file" ]]; then
        echo ""
        return
    fi
    
    # Extract the attribute value from the bztransmit tag
    grep -o "${attr}=\"[^\"]*\"" "$file" 2>/dev/null | sed "s/${attr}=\"\([^\"]*\)\"/\1/" || echo ""
}

# Check if overviewstatus.xml exists
if [[ ! -f "$OVERVIEWSTATUS_FILE" ]]; then
    log "ERROR: overviewstatus.xml not found at $OVERVIEWSTATUS_FILE"
    exit 1
fi

# Get current state and file
current_state=$(get_xml_attr "$OVERVIEWSTATUS_FILE" "cur_state")
current_file=$(get_xml_attr "$OVERVIEWSTATUS_FILE" "current_file")

log "Current state: '$current_state', Current file: '$current_file'"

# Check if we're in the problematic state
if [[ "$current_state" == "transmitting" ]] && [[ "$current_file" =~ ^ca[0-9]+/bz_done_[0-9]{8}_[0-9]+\.bzff$ ]]; then
    log "Detected potentially hanging transmission: $current_file"
    
    current_time=$(date +%s)
    
    # Check if we already have a state file
    if [[ -f "$STATE_FILE" ]]; then
        # Read the stored state
        read -r stored_state stored_file stored_timestamp < "$STATE_FILE"
        
        # Check if it's the same problematic state
        if [[ "$stored_state" == "$current_state" ]] && [[ "$stored_file" == "$current_file" ]]; then
            # Same hanging state - check how long it's been
            elapsed=$((current_time - stored_timestamp))
            log "Same hanging state detected for ${elapsed} seconds"
            
            if [[ $elapsed -ge $HANG_THRESHOLD_SECONDS ]]; then
                log "ERROR: Backup has been hanging for ${elapsed} seconds (threshold: ${HANG_THRESHOLD_SECONDS}s)"
                log "Triggering container restart by returning unhealthy status"
		kill -s 15 -1 && (sleep 10; kill -s 9 -1)
               # rm -f "$STATE_FILE"
                exit 1
            else
                remaining=$((HANG_THRESHOLD_SECONDS - elapsed))
                log "Monitoring hang state - ${remaining} seconds remaining before restart"
                exit 0
            fi
        else
            # Different state - update the state file
            log "Different hanging state detected - resetting timer"
            echo "$current_state $current_file $current_time" > "$STATE_FILE"
            exit 0
        fi
    else
        # First time seeing this state - create state file
        log "First detection of hanging state - starting timer"
        echo "$current_state $current_file $current_time" > "$STATE_FILE"
        exit 0
    fi
else
    # Not in problematic state - clean up state file if it exists
    if [[ -f "$STATE_FILE" ]]; then
        log "Not in hanging state - clearing previous hang detection"
        rm -f "$STATE_FILE"
    fi
    
    if [[ "$current_state" == "transmitting" ]]; then
        log "Transmitting normal file: $current_file"
    elif [[ "$current_state" == "not_running" ]]; then
        log "Backup not currently running"
    else
        log "Backup in state: $current_state"
    fi
    
    exit 0
fi

[mattlangis]

@JCStrainer how long did you wait? It can take some time until the email box appears the first time

[JCStrainer]

I fixed it by copying the installed files over from my windows setup. But I haven't found a solution for network shares, as such I have given up for now sadly.

Thank you for your script.

[truongty]

I got past the blue screen and am in the Backblaze Control Panel after rolling back to tessypowder/backblaze-personal-wine:v1.11

[sourcecodemage]

v11 seems to be working for me. I started with :latest which, based on the docker hub tag info, is probably v13 when I wrote this.

I am getting a warning to make sure that I have access to c:..\bzreports directory. I don't know if that is important to fix, but I'll look around the board for any reports about it.

[sourcecodemage]

I switched back to latest and it's working mostly fine. I noticed on the interface it seems to pause on a file, so there might be some container or backblaze application tweaks I need to do to improve performance.