Unable to launch Seafile 13 on macOS?

hieu · April 13, 2026, 1:16am

Unable to launch Seafile 13 on macOS via Docker Desktop. My docker compose file:

services:
seafile:
container_name: seafile
image: seafileltd/seafile-mc:13.0-latest
platform: linux/arm64/v8
restart: always
ports:

“9090:80”
volumes:
/Volumes/SeagateExos4x28TB_RAIDZ1/seafile/app:/shared
environment:
DB_HOST=seafile-mariadb
DB_ROOT_PASSWD=${SEAFILE_DB_ROOT_PASS}
TIME_ZONE=Asia/Ho_Chi_Minh
SEAFILE_ADMIN_EMAIL=${SEAFILE_ADMIN_EMAIL}
SEAFILE_ADMIN_PASSWORD=${SEAFILE_ADMIN_PASS}
SEAFILE_SERVER_LETSENCRYPT=false
SEAFILE_SERVER_HOSTNAME=seafile.doma.in
CACHE_PROVIDER=redis
REDIS_HOST=seafile-redis
REDIS_PORT=6379
REDIS_PASSWORD=${SEAFILE_REDIS_PASS}
NON_ROOT=true
extra_hosts:
“seafile.doma.in:${MY_SERVER_IP}”
“onlyoffice.doma.in:${MY_SERVER_IP}”
networks:
onlyoffice_office_network
depends_on:
seafile-mariadb:
condition: service_healthy
seafile-redis:
condition: service_started

seafile-mariadb:
container_name: seafile-mariadb
image: mariadb:10.11
platform: linux/arm64/v8
restart: always
environment:

MYSQL_ROOT_PASSWORD=${SEAFILE_DB_ROOT_PASS}
MYSQL_LOG_CONSOLE=true
volumes:
/Volumes/SeagateExos4x28TB_RAIDZ1/seafile/db:/var/lib/mysql
networks:
onlyoffice_office_network
healthcheck:
test: [“CMD”, “mariadb-admin”, “ping”, “-h”, “localhost”, “-u”, “root”, “-p${SEAFILE_DB_ROOT_PASS}”]
interval: 10s
timeout: 5s
retries: 5

seafile-redis:
container_name: seafile-redis
image: redis:7
platform: linux/arm64/v8
restart: always
command: redis-server --requirepass ${SEAFILE_REDIS_PASS}
volumes:

/Volumes/SeagateExos4x28TB_RAIDZ1/seafile/redis:/data
networks:
onlyoffice_office_network

networks:
onlyoffice_office_network:
external: true

volumes:
seafile_db_data:
seafile_redis_data:

Error log:

seafile | *** Running /etc/my_init.d/01_create_data_links.sh…
seafile | *** Booting runit daemon…
seafile | *** Runit started as PID 11
seafile | *** Running /scripts/enterpoint.sh…
seafile | 2026-04-13 02:17:29 Waiting Nginx
seafile | nginx: [warn] conflicting server name “” on 0.0.0.0:80, ignored
seafile | 2026-04-13 02:17:29 Nginx ready
seafile | 2026-04-13 02:17:29 Create linux user seafile in container, please wait.
seafile | 2026-04-13 02:17:29 The permission of path seafile/ is incorrect.
seafile | 2026-04-13 02:17:29 To use non root, change the folder permission of seafile folder in your host machine by ‘chmod -R a+rwx /opt/seafile-data/’ and try again later. (If you use another path, change the path in the command correspondingly). Now quit.
seafile | *** /scripts/enterpoint.sh exited with status 1.
seafile | *** Shutting down runit daemon (PID 11)…
seafile | *** Running /etc/my_init.post_shutdown.d/10_syslog-ng.shutdown…
seafile | *** Killing all processes…
seafile exited with code 1 (restarting)
seafile | *** Running /etc/my_init.d/01_create_data_links.sh…
seafile | *** Booting runit daemon…
seafile | *** Runit started as PID 11
seafile | *** Running /scripts/enterpoint.sh…
seafile | nginx: [warn] conflicting server name “” on 0.0.0.0:80, ignored
seafile | 2026-04-13 02:17:31 Nginx ready
seafile | 2026-04-13 02:17:31 Create linux user seafile in container, please wait.
seafile | groupadd: group ‘seafile’ already exists
seafile | useradd: user ‘seafile’ already exists
seafile | 2026-04-13 02:17:31 The permission of path seafile/ is incorrect.
seafile | 2026-04-13 02:17:31 To use non root, change the folder permission of seafile folder in your host machine by ‘chmod -R a+rwx /opt/seafile-data/’ and try again later. (If you use another path, change the path in the command correspondingly). Now quit.
seafile | *** /scripts/enterpoint.sh exited with status 1.
seafile | *** Shutting down runit daemon (PID 11)…
seafile | *** Running /etc/my_init.post_shutdown.d/10_syslog-ng.shutdown…
seafile | *** Killing all processes…
seafile exited with code 1 (restarting)
seafile | *** Running /etc/my_init.d/01_create_data_links.sh…
seafile | *** Booting runit daemon…
seafile | *** Runit started as PID 11
seafile | *** Running /scripts/enterpoint.sh…

daniel.pan · April 13, 2026, 8:55am

Here are suggestions from Our AI (manually checked) :

The error logs indicate that the container is failing because it cannot access the data directory with the necessary permissions required for non-root mode. When NON_ROOT=true is set, the Seafile container creates and runs as a specific user (UID 8000), and it currently lacks write access to your external volume.

According to the Seafile Admin Manual , you should follow these steps to resolve the issue:

1. Fix Permissions on the Host

The most direct fix is to grant the necessary permissions to the folder on your macOS host. Open your terminal and run:

sudo chmod -R a+rwx /Volumes/SeagateExos4x28TB_RAIDZ1/seafile/app

2. Initialize as Root First

If this is a new installation, the initialization process often requires root privileges to set up the internal folder structure correctly before switching to a non-root user.

In your docker-compose.yml, set NON_ROOT=false.
Run docker compose up -d and wait for the system to initialize completely.
Stop and remove the containers: docker compose down.
Apply the chmod command mentioned in Step 1 to your host directory.
Change back to NON_ROOT=true in your configuration and restart: docker compose up -d .

3. Check macOS File Sharing

Ensure that the path /Volumes/SeagateExos4x28TB_RAIDZ1/ is permitted in your Docker Desktop settings:

Go to Settings > Resources > File Sharing.
Add your external drive path to the list and click Apply & Restart.

4. Filesystem Considerations

Since you are using a RAIDZ1 volume on macOS, ensure the underlying filesystem supports POSIX permissions. If the drive is formatted as exFAT, it may not handle Linux-style UID/GID permissions correctly. In such cases, running with NON_ROOT=false might be the only stable workaround for that specific hardware setup.

hieu · April 13, 2026, 10:11am

I followed those steps. However, I am still unable to launch Seafile 13. I used to use Seafile 11, and Seafile 12 worked fine.

error log from docker desktop:

seafile-redis | 2026-04-13 17:04:19.624 | 1:C 13 Apr 2026 10:04:19.624 * oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
seafile-redis | 2026-04-13 17:04:19.624 | 1:C 13 Apr 2026 10:04:19.624 * Redis version=7.4.8, bits=64, commit=00000000, modified=0, pid=1, just started
seafile-redis | 2026-04-13 17:04:19.624 | 1:C 13 Apr 2026 10:04:19.624 * Configuration loaded
seafile-redis | 2026-04-13 17:04:19.624 | 1:M 13 Apr 2026 10:04:19.624 * monotonic clock: POSIX clock_gettime
seafile-redis | 2026-04-13 17:04:19.625 | 1:M 13 Apr 2026 10:04:19.625 * Running mode=standalone, port=6379.
seafile-redis | 2026-04-13 17:04:19.626 | 1:M 13 Apr 2026 10:04:19.625 * Server initialized
seafile-redis | 2026-04-13 17:04:19.628 | 1:M 13 Apr 2026 10:04:19.627 * Loading RDB produced by version 7.4.8
seafile-redis | 2026-04-13 17:04:19.628 | 1:M 13 Apr 2026 10:04:19.627 * RDB age 66 seconds
seafile-redis | 2026-04-13 17:04:19.628 | 1:M 13 Apr 2026 10:04:19.627 * RDB memory usage when created 0.94 Mb
seafile-redis | 2026-04-13 17:04:19.628 | 1:M 13 Apr 2026 10:04:19.627 * Done loading RDB, keys loaded: 0, keys expired: 0.
seafile-redis | 2026-04-13 17:04:19.628 | 1:M 13 Apr 2026 10:04:19.628 * DB loaded from disk: 0.002 seconds
seafile-redis | 2026-04-13 17:04:19.628 | 1:M 13 Apr 2026 10:04:19.628 * Ready to accept connections tcp
seafile-mariadb | 2026-04-13 17:04:19.667 | 2026-04-13 10:04:19+00:00 [Note] [Entrypoint]: Entrypoint script for MariaDB Server 1:10.11.16+maria~ubu2204 started.
seafile-mariadb | 2026-04-13 17:04:19.911 | 2026-04-13 10:04:19+00:00 [Warn] [Entrypoint]: /sys/fs/cgroup///memory.pressure not writable, functionality unavailable to MariaDB
seafile-mariadb | 2026-04-13 17:04:19.912 | 2026-04-13 10:04:19+00:00 [Note] [Entrypoint]: Switching to dedicated user ‘mysql’
seafile-mariadb | 2026-04-13 17:04:19.916 | 2026-04-13 10:04:19+00:00 [Note] [Entrypoint]: Entrypoint script for MariaDB Server 1:10.11.16+maria~ubu2204 started.
seafile-mariadb | 2026-04-13 17:04:20.060 | 2026-04-13 10:04:20+00:00 [Note] [Entrypoint]: MariaDB upgrade not required
seafile-mariadb | 2026-04-13 17:04:20.081 | 2026-04-13 10:04:20 0 [Note] Starting MariaDB 10.11.16-MariaDB-ubu2204 source revision 3218602d3100db9ce7a875511a591cddc173cc16 server_uid aSuB9hLjN8V5fw6LlX93dX/OfGg= as process 1
seafile-mariadb | 2026-04-13 17:04:20.102 | 2026-04-13 10:04:20 0 [Note] InnoDB: Compressed tables use zlib 1.2.11
seafile-mariadb | 2026-04-13 17:04:20.103 | 2026-04-13 10:04:20 0 [Note] InnoDB: Number of transaction pools: 1
seafile-mariadb | 2026-04-13 17:04:20.103 | 2026-04-13 10:04:20 0 [Note] InnoDB: Using ARMv8 crc32 + pmull instructions
seafile-mariadb | 2026-04-13 17:04:20.107 | 2026-04-13 10:04:20 0 [Warning] mariadbd: io_uring_queue_init() failed with EPERM: sysctl kernel.io_uring_disabled has the value 2, or 1 and the user of the process is not a member of sysctl kernel.io_uring_group. (see man 2 io_uring_setup).
seafile-mariadb | 2026-04-13 17:04:20.107 | create_uring failed: falling back to libaio
seafile-mariadb | 2026-04-13 17:04:20.107 | 2026-04-13 10:04:20 0 [Note] InnoDB: Using Linux native AIO
seafile-mariadb | 2026-04-13 17:04:20.108 | 2026-04-13 10:04:20 0 [Note] InnoDB: innodb_buffer_pool_size_max=128m, innodb_buffer_pool_size=128m
seafile-mariadb | 2026-04-13 17:04:20.108 | 2026-04-13 10:04:20 0 [Note] InnoDB: Completed initialization of buffer pool
seafile-mariadb | 2026-04-13 17:04:20.110 | 2026-04-13 10:04:20 0 [Note] InnoDB: Buffered log writes (block size=512 bytes)
seafile-mariadb | 2026-04-13 17:04:20.144 | 2026-04-13 10:04:20 0 [Note] InnoDB: End of log at LSN=44808
seafile-mariadb | 2026-04-13 17:04:20.182 | 2026-04-13 10:04:20 0 [Note] InnoDB: 128 rollback segments are active.
seafile-mariadb | 2026-04-13 17:04:20.183 | 2026-04-13 10:04:20 0 [Note] InnoDB: Setting file ‘./ibtmp1’ size to 12.000MiB. Physically writing the file full; Please wait …
seafile-mariadb | 2026-04-13 17:04:20.183 | 2026-04-13 10:04:20 0 [Note] InnoDB: File ‘./ibtmp1’ size is now 12.000MiB.
seafile-mariadb | 2026-04-13 17:04:20.184 | 2026-04-13 10:04:20 0 [Note] InnoDB: log sequence number 44808; transaction id 14
seafile-mariadb | 2026-04-13 17:04:20.184 | 2026-04-13 10:04:20 0 [Note] Plugin ‘FEEDBACK’ is disabled.
seafile-mariadb | 2026-04-13 17:04:20.185 | 2026-04-13 10:04:20 0 [Note] InnoDB: Loading buffer pool(s) from /var/lib/mysql/ib_buffer_pool
seafile-mariadb | 2026-04-13 17:04:20.187 | 2026-04-13 10:04:20 0 [Warning] You need to use --log-bin to make --expire-logs-days or --binlog-expire-logs-seconds work.
seafile-mariadb | 2026-04-13 17:04:20.190 | 2026-04-13 10:04:20 0 [Note] InnoDB: Buffer pool(s) load completed at 260413 10:04:20
seafile-mariadb | 2026-04-13 17:04:20.191 | 2026-04-13 10:04:20 0 [Note] Server socket created on IP: ‘0.0.0.0’, port: ‘3306’.
seafile-mariadb | 2026-04-13 17:04:20.191 | 2026-04-13 10:04:20 0 [Note] Server socket created on IP: ‘::’, port: ‘3306’.
seafile-mariadb | 2026-04-13 17:04:20.224 | 2026-04-13 10:04:20 0 [Note] mariadbd: ready for connections.
seafile-mariadb | 2026-04-13 17:04:20.224 | Version: ‘10.11.16-MariaDB-ubu2204’ socket: ‘/run/mysqld/mysqld.sock’ port: 3306 mariadb. org binary distribution
seafile | 2026-04-13 17:04:30.724 | *** Running /etc/my_init.d/01_create_data_links.sh…
seafile | 2026-04-13 17:04:30.734 | *** Booting runit daemon…
seafile | 2026-04-13 17:04:30.734 | *** Runit started as PID 11
seafile | 2026-04-13 17:04:30.734 | *** Running /scripts/enterpoint.sh…
seafile | 2026-04-13 17:04:30.739 | 2026-04-13 17:04:30 Waiting Nginx
seafile | 2026-04-13 17:04:30.750 | nginx: [warn] conflicting server name “” on 0.0.0.0:80, ignored
seafile | 2026-04-13 17:04:30.946 | 2026-04-13 17:04:30 Nginx ready
seafile | 2026-04-13 17:04:30.947 | 2026-04-13 17:04:30 Create linux user seafile in container, please wait.
seafile | 2026-04-13 17:04:47.487 | 2026-04-13 17:04:47 This is an idle script (infinite loop) to keep container running.
seafile | 2026-04-13 17:04:47.617 | nginx: [warn] conflicting server name “” on 0.0.0.0:80, ignored

-enterpoint.log-

[2026-04-13 17:02:52] Waiting Nginx
[2026-04-13 17:02:52] Nginx ready
[2026-04-13 17:02:52] This is an idle script (infinite loop) to keep container running.
[2026-04-13 17:04:30] Waiting Nginx
[2026-04-13 17:04:30] Nginx ready
[2026-04-13 17:04:30] Create linux user seafile in container, please wait.
[2026-04-13 17:04:47] This is an idle script (infinite loop) to keep container running.

daniel.pan · April 13, 2026, 10:35am

From your latest log of docker, the services are started just fine.

What othere errors do you have? Can you find some errors in the log files?

hieu · April 13, 2026, 12:32pm

When I try to access Seafile 13 via my local IP:9090 on the same host machine, it shows a 502 Bad Gateway error.

daniel.pan · April 14, 2026, 12:15am

The “502 Bad Gateway” error confirms that your Nginx container is working, but it cannot reach the Seahub service (running on port 8000 internally). In Seafile 13, this is often a sign that Seahub failed to start.

Since you are using an external RAIDZ1 volume on macOS with NON_ROOT=true, the most likely cause is that the internal seafile user (UID 8000) still cannot write to specific sub-directories required for Seahub to start (like logs or pids).

Please try the following steps to diagnose and fix this:

Verify Seahub Logs:
Check the internal Seahub logs for specific Python errors. You can find them on your host at:
/Volumes/SeagateExos4x28TB_RAIDZ1/seafile/app/seafile/logs/seahub.log
If you see “Permission denied” errors here, the NON_ROOT mode is still struggling with the macOS filesystem mapping.
Test with Root Mode:
To rule out macOS filesystem/UID mapping issues, temporarily set NON_ROOT=false in your docker-compose.yml, run docker compose up -d, and see if the 502 error clears. If it works as root, the issue is strictly how macOS handles permissions on your external RAID volume.

hieu · April 14, 2026, 9:22am

Even after setting NON_ROOT=false, I’m still seeing the ‘502 Bad Gateway’ error.

–seahub.error.log-

2026/04/14 16:11:02 [error] 64#64: *1 connect() failed (111: Connection refused) while connecting to upstream, client: 172.65.xx.xxx, server: , request: “GET /api2/repos/ HTTP/1.1”, upstream: “http:// 127.0.0.1:8000/api2/repos/”, host: “seafile.doma.in”

-seahub.access.log-

172.65.xx.xxx 172.65.xx.xxx [14/Apr/2026:16:10:28 +0700] “GET /api2/repos/ HTTP/1.1” 502 150 “-” “Mozilla/5.0” 0.001

- 172.65.xx.xxx [14/Apr/2026:16:10:29 +0700] “GET / HTTP/1.1” 502 552 “-” “Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/146.0.0.0 Safari/537.36” 0.000

- 172.65.xx.xxx [14/Apr/2026:16:10:29 +0700] “GET /favicon.ico HTTP/1.1” 502 552 “http:// 192.168.1.205:9090/” “Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/146.0.0.0 Safari/537.36” 0.000

172.65.xx.xxx 172.65.xx.xxx [14/Apr/2026:16:10:32 +0700] “GET /api2/repos/ HTTP/1.1” 502 150 “-” “Mozilla/5.0” 0.000

daniel.pan · April 14, 2026, 9:29am

To further debug the “502 Bad Gateway” error, we need to see the actual error message Seahub produces when it attempts to start. In the current setup, Seahub runs in “daemon” mode (in the background), which often swallows the initial startup errors.

I suggest you manually start Seahub in the foreground to catch the error:

Enter the Seafile container:
```
docker exec -it seafile /bin/bash
```
Modify Gunicorn settings:
Navigate to your configuration directory (usually /shared/seafile/conf) and edit gunicorn.conf.py. Find the line that says daemon = True and change it to:
```
daemon = False
```
Attempt to start Seahub manually:
From within the container (usually in /opt/seafile/seafile-server-latest), run:
```
./seahub.sh start
```

By running it this way, Seahub will not return to the command prompt. Instead, it will stay active in your terminal and print the full Python traceback or any “Permission denied” errors directly to your screen. This should reveal exactly why the service is failing to bind to port 8000.

Once you find the error, you can revert the daemon setting back to True and restart the container.

hieu · April 14, 2026, 10:27am

gunicorn.conf.py might not have been created yet?

daniel.pan · April 14, 2026, 11:29pm

The path should be /opt/seafile/conf/gunicorn.conf.py. If the file is not generated, your initialization process is not correctly finished.

I rechecked your inital post, it seems that you have modified the yml files. According to seafile setup manual, you should not modify the yml files directly, only modify the environment file .env. There likely to be some errors after you modify yml files.

hieu · April 15, 2026, 1:32pm

I don’t think I have changed the YAML files. I have deleted everything under /Volumes/SeagateExos4x28TB_RAIDZ1/seafile/*.

Then, run the compose file as described in my initial post with NON_ROOT=false.

but gunicorn.conf.py hasn’t been generated in the path.