Open nkgiovannivl opened 1 year ago
I'm having the same issue.
Same here. Any solution for this available yet?
Having the same problem. Will there be fix for this in the near future?
Ill try this config later this week to see if I can reproduce.
I second that. Same problem for me. I hope somebody else fix it. I've tried using UFW to allow ports. Nothing seems to work.
Using the following docker-compose.yml worked for me. Obsidian-remote now works, but I don't like how it works. It basically uses Guacamole as VNC server/viewer. It doesn't fit the whole browser window if you click on it.
Docker-compose.yml config I found (you can modify it according to your own taste): version: "3.9" services: obsidian: image: ghcr.io/sytone/obsidian-remote:latest container_name: Obsidian-Remote hostname: obsidian-remote mem_limit: 1g cpu_shares: 768 security_opt:
same problem here, use shell interact with the container find that the server failed starting.
ghcr.io/sytone/obsidian-remote:latest
is outdated with the last release of it 9 months ago. Changes to the base image and some fixes have went out since - @sytone can we get an update to the Docker tag to better help you debug this on the current codebase?
I solved it by removing the "" from the port, like so
- CUSTOM_PORT=8080
- CUSTOM_HTTPS_PORT=8443
I solved it by removing the "" from the port, like so
- CUSTOM_PORT=8080 - CUSTOM_HTTPS_PORT=8443
Hi there! I tried doing what you did but mine just said websocat: invalid port number
. I'm doing the installation through Codesandbox for context.
Hi there! I tried doing what you did but mine just said
websocat: invalid port number
. I'm doing the installation through Codesandbox for context.
I'm getting the same message and when loading the page in a browser I also get Cannot GET /
.
This is the state of my docker-compose.
services:
obsidian:
image: 'ghcr.io/sytone/obsidian-remote:latest'
container_name: obsidian-remote
restart: unless-stopped
ports:
- 8081:8081
- 8443:8443
volumes:
- /home/obsidian/vaults:/vaults
- /home/obsidian/config:/config
environment:
- PUID=1000
- PGID=1000
- TZ=Europe/Stockholm
- DOCKER_MODS=linuxserver/mods:universal-git
- CUSTOM_PORT=8081
- CUSTOM_HTTPS_PORT=8443
- CUSTOM_USER=""
- PASSWORD=""
- SUBFOLDER=""
Hi Guys,
By doing a bit of a mix of the two docker compose that are in this thread here is one that works:
obsidian:
image: 'ghcr.io/sytone/obsidian-remote:latest'
container_name: obsidian-remote
restart: unless-stopped
hostname: obsidian-remote
mem_limit: 2g
cpu_shares: 768
security_opt:
- no-new-privileges:true
healthcheck:
test: curl -f http://localhost:8080/ || exit 1
ports:
- 8080:8080
- 8443:8443
volumes:
- /home/tosmaniac/docker/obsidian/vaults:/vaults
- /home/tosmaniac/docker/obsidian/config:/config
environment:
- PUID=1000
- PGID=100
- TZ=European/Berlin
- DOCKER_MODS=linuxserver/mods:universal-git
# - CUSTOM_PORT="8080"
# - CUSTOM_HTTPS_PORT="8443"
# - CUSTOM_USER=""
# - PASSWORD=""
# - SUBFOLDER=""```
Latest version of obsidian gets installed.
Thank you all.
Florent
It seems like guacd
closing the connection during the handshake with guacamole-lite
(I think it runs both? unsure to be honest, I've never used guacamole before) because of some improperly mapped configuration files. They exist in the image, but I think unpinned versions caused it to break after initial release.
Looking into the container with docker compose exec -it obsidian-remote /bin/bash
I can see that guaclog
outputs the following:
root@obsidian-remote:/# guaclog
guaclog: INFO: Guacamole input log interpreter (guaclog) version 1.1.0
guaclog: INFO: No input files specified. Nothing to do.
I also noticed that the user and group within the container is abc:abc
, potentially incorrect user mapping is to blame.
I stopped at the point where I'd have to clone the repo and build the image from the Dockerfile as I am running out of time to fiddle with it today, if anyone gets it working definitely let me know.
p.s. the docker compose config from @tosmaniac modified to have the commented fields uncommented and port numbers unquoted was the closest I managed to get, only outputting this error instead of the longer ones above.
obsidian-remote | guacd[208]: INFO: Guacamole proxy daemon (guacd) version 1.1.0 started
obsidian-remote | guacd[208]: INFO: Listening on host 0.0.0.0, port 4822
obsidian-remote | Starting guacamole-lite websocket server
obsidian-remote | listening on *:8080
obsidian-remote | guacd[208]: INFO: Guacamole connection closed during handshake
obsidian-remote | [guac-init] Auto start set, starting application
obsidian-remote | websocat: invalid port number
I had the websocat: invalid port number
, but got it working. I don't understand the fix, but I started removing arguments to get to basic config until I got it working.
Key adjustment steps as I progressed:
SUBFOLDER
env variable. An empty string causes problemswebsocat
error was to remove CUSTOM-USER
AND PASSWORD
env variables. I just switched to auth through my reverse proxy, instead. Here's my final docker-compose: obsidian:
container_name: obsidian
image: 'ghcr.io/sytone/obsidian-remote:latest'
restart: unless-stopped
ports:
- 8080:8080
volumes:
- ./obsidian/vaults:/vaults
- ./obsidian/config:/config
environment:
- PUID=1000
- PGID=1000
- TZ=America/New_York
- DOCKER_MODS=linuxserver/mods:universal-git
I ended up having to do a number things, which leads me to believe there's not one issue, but several small issues.
This gave me a little bit of progress, but I couldn't yet open Obsidian successfully. I got some sort of error like /squashfs-root/obsidian is not a directory
.
To fix this issue, I modified some of the files that were created within the config directory. (mine is setup to the my-config
folder)
The last change is not required, but something I did just to improve my QoL (Qualify of Life).
Is this on AMD or ARM?
obsidian:
image: 'ghcr.io/sytone/obsidian-remote:latest'
container_name: obsidian-remote
restart: unless-stopped
ports:
- 6666:8080
- 8663:8443
volumes:
- ${DATA}/obsidian/vaults:/vaults
- ${DATA}/obsidian/config:/config
environment:
- PUID=1003
- PGID=1003
- TZ=${TZ}
- DOCKER_MODS=linuxserver/mods:universal-git
- CUSTOM_PORT=8080
- CUSTOM_HTTPS_PORT=8443
- CUSTOM_USER=123
- PASSWORD=123
with this config i got no websocat error, but in the logs i got this:
[2024-02-06 11:28:48] [Connection 1] Closing connection with error: Error: WS was inactive for too long
at ClientConnection.checkActivity (/gclient/node_modules/guacamole-lite/lib/ClientConnection.js:154:24)
at listOnTimeout (internal/timers.js:557:17)
[2024-02-06 11:28:48] [Connection 1] Closing guacd connection
at processTimers (internal/timers.js:500:7)
[2024-02-06 11:28:48] [Connection 1] Client connection closed
guacd[1135]: ERROR: User is not responding.
guacd[1135]: INFO: User "@085c8313-50ce-4eb0-872a-05e831788aae" disconnected (0 users remain)
guacd[1135]: INFO: Last user of connection "$9225c7a9-a824-415c-8041-03bcd513541e" disconnected
guacd[1135]: INFO: Internal RDP client disconnected
guacd[1100]: INFO: Connection "$9225c7a9-a824-415c-8041-03bcd513541e" removed.
rdpClientConRecv: g_sck_recv failed(returned 0)
rdpClientConRecvMsg: error
rdpClientConCheck: rdpClientConGotData failed
rdpClientConDisconnect:
rdpRemoveClientConFromDev: removing clientCon 0x5617e058a850
it doesnt matter if i use the last four lines with the costum ports, user and passwort, same result.
does anybody got it working?
obsidian: image: 'ghcr.io/sytone/obsidian-remote:latest' container_name: obsidian-remote restart: unless-stopped ports: - 6666:8080 - 8663:8443 volumes: - ${DATA}/obsidian/vaults:/vaults - ${DATA}/obsidian/config:/config environment: - PUID=1003 - PGID=1003 - TZ=${TZ} - DOCKER_MODS=linuxserver/mods:universal-git - CUSTOM_PORT=8080 - CUSTOM_HTTPS_PORT=8443 - CUSTOM_USER=123 - PASSWORD=123
with this config i got no websocat error, but in the logs i got this:
[2024-02-06 11:28:48] [Connection 1] Closing connection with error: Error: WS was inactive for too long at ClientConnection.checkActivity (/gclient/node_modules/guacamole-lite/lib/ClientConnection.js:154:24) at listOnTimeout (internal/timers.js:557:17) [2024-02-06 11:28:48] [Connection 1] Closing guacd connection at processTimers (internal/timers.js:500:7) [2024-02-06 11:28:48] [Connection 1] Client connection closed guacd[1135]: ERROR: User is not responding. guacd[1135]: INFO: User "@085c8313-50ce-4eb0-872a-05e831788aae" disconnected (0 users remain) guacd[1135]: INFO: Last user of connection "$9225c7a9-a824-415c-8041-03bcd513541e" disconnected guacd[1135]: INFO: Internal RDP client disconnected guacd[1100]: INFO: Connection "$9225c7a9-a824-415c-8041-03bcd513541e" removed. rdpClientConRecv: g_sck_recv failed(returned 0) rdpClientConRecvMsg: error rdpClientConCheck: rdpClientConGotData failed rdpClientConDisconnect: rdpRemoveClientConFromDev: removing clientCon 0x5617e058a850
it doesnt matter if i use the last four lines with the costum ports, user and passwort, same result.
does anybody got it working?
I'm having the same issue, Tried all suggested changes above but nothing is working for me.
I was able to get it working by pulling this newer image not from ghcr.io
services:
obsidian:
image: sytone/obsidian-remote:latest
container_name: obsidian
hostname: obsidian
restart: unless-stopped
ports:
- 8080:8080
- 8443:8443
volumes:
- /obsidian/vaults:/vaults
- /obsidian/config:/config
environment:
- PUID=1000
- PGID=1000
- TZ=America/Toronto
- DOCKER_MODS=linuxserver/mods:universal-git```
In 2024 images from ghcr.io are outdated. It is better to use images from dockerhub with latest Obsidian installations.
Folders involved:
/home/user/Documents/obsidian
- obsidian Vault named "obsidian" - folder for my local installation, with installed plugins, .obsidian dir, md files, excalidraw, etc
/home/user/Documents/obsidian-remote
- some new empty folder - used as docker volume shared folder
Execute:
mkdir -p /home/user/Documents/obsidian-remote/vaults
cp -R /home/user/Documents/obsidian /home/user/Documents/obsidian-remote/vaults/obsidian
- copy Vault named "obsidian" to docker volume shared folderAttention:
PUID,PGID
in the docker-compose
I had problems running container with my host machine user id&group (1000). I was forced to set these environment variables as root (0).
Please make attention to file permissions inside obsidian-remote
host folder. It is better to sudo rm -r
it and make a fresh copy of your local Vault folder
After docker-compose has been started, please open http://localhost:8082/ in your browser. If there is kind of black screen or empty desktop:
cd /opt/obsidian
./obsidian --no-sandbox
Open - Computer
, /vaults/obsidian
version: '3.8'
services:
obsidian:
image: 'sytone/obsidian-remote:latest'
container_name: obsidian-remote
restart: unless-stopped
hostname: obsidian-remote
security_opt:
- no-new-privileges:true
healthcheck:
test: curl -f http://localhost:8080/ || exit 1
ports:
- 8082:8080
- 8445:8443
volumes:
- /home/user/Documents/obsidian-remote/vaults:/vaults
- /home/user/Documents/obsidian-remote/config:/config
environment:
- PUID=0
- PGID=0
- TZ=European/Kyiv
- DOCKER_MODS=linuxserver/mods:universal-git
In 2024 images from ghcr.io are outdated. It is better to use images from dockerhub with latest Obsidian installations.
Preparation
Folders involved:
/home/user/Documents/obsidian
- obsidian Vault named "obsidian" - folder for my local installation, with installed plugins, .obsidian dir, md files, excalidraw, etc/home/user/Documents/obsidian-remote
- some new empty folder - used as docker volume shared folderExecute:
1. `mkdir -p /home/user/Documents/obsidian-remote/vaults` 2. `cp -R /home/user/Documents/obsidian /home/user/Documents/obsidian-remote/vaults/obsidian` - copy Vault named "obsidian" to docker volume shared folder
Attention:
PUID,PGID
inthe docker-compose
I had problems running container with my host machine user id&group (1000). I was forced to set these environment variables as root (0). Please make attention to file permissions insideobsidian-remote
host folder. It is better tosudo rm -r
it and make a fresh copy of your local Vault folderAfter docker-compose has been started, please open http://localhost:8082/ in your browser. If there is kind of black screen or empty desktop:
1. right-click - select xterm 2. `cd /opt/obsidian` 3. `./obsidian --no-sandbox` * in Obsidian you can open existing Vault `Open - Computer`, `/vaults/obsidian` * click "maximize" window button to maximize Obsidian
docker-compose.yml
version: '3.8' services: obsidian: image: 'sytone/obsidian-remote:latest' container_name: obsidian-remote restart: unless-stopped hostname: obsidian-remote security_opt: - no-new-privileges:true healthcheck: test: curl -f http://localhost:8080/ || exit 1 ports: - 8082:8080 - 8445:8443 volumes: - /home/user/Documents/obsidian-remote/vaults:/vaults - /home/user/Documents/obsidian-remote/config:/config environment: - PUID=0 - PGID=0 - TZ=European/Kyiv - DOCKER_MODS=linuxserver/mods:universal-git
The following below does not work for me when I have a black screen. The directory does not exist.
After docker-compose has been started, please open http://localhost:8082/ in your browser. If there is kind of black screen or empty desktop:
right-click - select xterm
cd /opt/obsidian
./obsidian --no-sandbox
@2ndwonder This is why you need root privileges. Otherwise you cannot access the system files /opt/obsidian
First, I am using the one from linuxserver.io, as this project did not seem to work for me, ( white window)
In portainer terminal it was giving error it cannot open the /squashfs-root/obsidian
, ofcourse not as there is no folder like that in root of the container.
Further checking where the obsidian is on system from the remote terminal which obsidian
it is in the /usr/bin/obsidian
checking further, there is script to start obsidian, for some reason using the squashfs-root folder, so i just change it to correct location and it works...
nano /config/.config/openbox/autostart
# Comment out the original line and put the good path bellow
#/squashfs-root/obsidian --no-sandbox --disable-dev-shm-usage --disable-gpu --disable-software-rasterizer
/usr/bin/obsidian --no-sandbox --disable-dev-shm-usage --disable-gpu --disable-software-rasterizer
the config folder should be somewhere on your system, as it is maped via docker-compose (in my case)
Describe the bug At first setup of docker compose logs shows error related to
Guacamole connection closed during handshake
To Reproduce Below docker compose:
Log info from docker:
Expected behavior Start normally
Screenshots N/A
Desktop (please complete the following information):
N/A
Smartphone (please complete the following information):
N/A
Additional context Add any other context about the problem here.