So I am running a Nextcloud 20 setup using docker-compose () Nextcloud working perfect, pointing the settings for Collabora office within Nextcloud gives me a green sign. I had this working with the same setup running NC 18.
I am using Swag as a reverse proxy, both pointing to the nextcloud containers.
But when trying to open a document - it just becomes a blank page.
Going to the Collabora URL gives me an simple - OK.
I am lost - what am I doing wrong here?
docker-compose-file:
version: ‘3.7’
services:
swag:
image: linuxserver/swag:latest
container_name: swag
cap_add:
- NET_ADMIN
environment:
- PUID=1000
- PGID=1000
- TZ=Europe/ducktown
- URL=clouddomain.io
- SUBDOMAINS=wildcard
- VALIDATION=dns
- DNSPLUGIN=cloudflare
- EMAIL=my.own@domian.io
- STAGING=false
volumes:
- ./appdata/swag:/config
ports:
- 443:443
- 80:80
restart: unless-stopped
nextcloud:
image: linuxserver/nextcloud:latest
container_name: nextcloud
hostname: cloud
environment:
- PUID=1000
- PGID=1000
- TZ=Europe/Ducktown
volumes:
- ./appdata/nextcloud:/config
- Cloud:/data (An NFS mounted file array)
depends_on:
- mariadb
restart: unless-stopped
mariadb:
image: linuxserver/mariadb:latest
container_name: mariadb
environment:
- PUID=1000
- PGID=1000
- MYSQL_ROOT_PASSWORD=a-super-secret-passsword
- TZ=Europe/ducktown
volumes:
- ./appdata/mariadb:/config
ports:
- 3306:3306
restart: unless-stopped
collabora:
image: collabora/code:latest
container_name: collabora
hostname: office
environment:
- domain=office\.clouddomain\.io
- username=officeadmin
- password=another-super-secret-password
restart: unless-stopped
volumes:
- ./appdata/collabora:/config
ports:
- 9980:9980
Hello and thanks for reporting this!
I’m not that versed in this subject but if you can reach it (you do see the plain page with “OK”) it seems that you are getting the bare url (without the long addition starting with /loleaflet) when opening a document, and so it does not open any particular app (writer, calc, impress) but instead gives you the generic status page.
maybe someone can chime in with a more insightful answer.
The ok comes from when I visit the URL without opening a document.
Like https://office.domain.io
I have been asking around to see if someone has a solution but the response I scarse.
I have been playing around with different settings for the Collabora container.
collabora:
image: collabora/code:latest
container_name: collabora
hostname: office
environment:
- domain=office.domain.io
- username=officeadmin
- password=
- dictionaries=en
- extra_params=–o:ssl.enable=false --o:ssl.termination=true
restart: unless-stopped
volumes:
- ./appdata/collabora:/config
ports:
- 9980:9980
But for some reason noting works for me, thing is I had it working with Nextcloud 18.
I really do hope someone can point me in the right direction.
I got some help but it doesn’t seem to be my problem by interesting nonetheless:
“When I opened up my JavaScript console I noticed that Chrome was blocking the request to Collabora since Nextcloud was running over HTTPS but the request to Collabora was being made over HTTP. This was odd since I setup both to be reverse proxied behind Nginx, and I had setup Collabora with an HTTPS address in the Nextcloud config page.”
@mattias interesting indeed. I have share your topic in our community channel* : ) let’s seem if someone can help you.
*
Please do. I have been digging deep into this, still the same issue.
I am open to show everything to anyone that can help me.
But I am now at the level of digging through logs, as one does to solve issues.
I found this interesting part:
yesterday at 2:29 PM kit-00045-00039 2021-01-08 13:29:54.166601 [ kit_spare_002 ] ERR Failed to get the realpath of [/etc/timezone] (ENOENT: No such file or directory)| common/FileUtil.cpp:256
yesterday at 2:29 PM kit-00045-00039 2021-01-08 13:29:54.166820 [ kit_spare_002 ] ERR Failed to get the realpath of [/etc/localtime] (ENOENT: No such file or directory)| common/FileUtil.cpp:256
yesterday at 3:07 PM wsd-00007-00043 2021-01-08 14:07:02.250919 [ websrv_poll ] WRN Poco::Net::DNS::resolve(“10.100.100.1”) failed: Host not found: 10.100.100.1| wsd/LOOLWSD.cpp:2327
yesterday at 3:07 PM wsd-00007-00043 2021-01-08 14:07:02.251445 [ websrv_poll ] WRN convert-to: Requesting address is denied: 10.100.100.1| wsd/LOOLWSD.cpp:2333
yesterday at 3:07 PM wsd-00007-00043 2021-01-08 14:07:21.870591 [ websrv_poll ] ERR Socket #26 SSL BIO error: closed unexpectedly (-1). (0: Success)| ./net/SslSocket.hpp:276
yesterday at 3:07 PM wsd-00007-00043 2021-01-08 14:07:21.871366 [ websrv_poll ] ERR Error while handling poll for socket #26 in websrv_poll: SSL Socket closed unexpectedly.| net/Socket.cpp:312
yesterday at 3:07 PM wsd-00007-00043 2021-01-08 14:07:28.971695 [ websrv_poll ] WRN Poco::Net::DNS::resolve(“10.100.100.1”) failed: Host not found: 10.100.100.1| wsd/LOOLWSD.cpp:2327
yesterday at 3:07 PM wsd-00007-00043 2021-01-08 14:07:28.972022 [ websrv_poll ] WRN convert-to: Requesting address is denied: 10.100.100.1| wsd/LOOLWSD.cpp:2333
There are 2 things that stands out more one than the other.
First: Failed to get the realpath of [/etc/timezone]
And then: wsd-00007-00043 2021-01-08 14:07:02.250919 [ websrv_poll ] WRN Poco::Net::DNS::resolve(“10.100.100.1”) failed: Host not found: 10.100.100.1| wsd/LOOLWSD.cpp:2327
Also when I do a docker-compose down, and then a docker ps and Collabora container is still up…
That one one has no image in docker images, still up and running.
So during the day I have solved the issue, it seems to be a code change from Collaboras side that breaks my setup.
First off I changed the - domain = in the Collabora environment from its own to the same a nextclouds.
Secondly I tested and have come to the conclusion that collabora/code:6.4.1.3 is the latest I can use, it is when the version goes to 6.4.2 that my setup breaks.
so I have been looking and asking around and a PR that fixes that is already pending and it will soon be merged to Master.
Thanks for all the info and logs and of course for reporting it!
1 Like