I have been running a workspace with document server on my digital ocean ubuntu server droplet for over a year and 2 or 3 times a week, the CRM will crash and the server will need a restart. Today I noticed a docker update (after a crash) and foolishly run the update midday, before backing up.
We got the stuck loading screen and couldnt get passed it. So I consulted a LLM (perhaps another foolish decision) to try and troubleshoot and ended up backing up my database and re-installing the onlyoffice docker. The only thing I changed from past installs was not including the mailserver module because for some reason it wouldnt accept my domain.
Unfortunately, I have been stuck in a 8 hour loop of failed attempts to get things running. The best guess from the LLM is that the community server is not able to load and it is likely because of a silent failure due to incompatibility with the backed up database.
Here is a direct quote:
Final Troubleshooting and Fix for OnlyOffice CRM Stuck LoadingI understand the frustration — the logs stopping at “success: www entered RUNNING state” means nginx is up, but the OnlyOffice backend (the .NET Core app for the portal and CRM) is crashing silently during initialization, preventing migrations and the “Server started” line. This is a known issue in OnlyOffice Docker setups after updates or reinstalls (from the web search results, e.g., where the community-server container restarts endlessly due to update failure, where the portal hangs on startup after Docker install, and where the container crashes after 6 seconds with no UI, fixed by full reinstall or dependency fixes). The curl 302 is nginx responding, but the app backend never loads, so the portal stays stuck on the loading bar, and CRM API endpoints fail.The root cause is likely:MySQL connection failure: The app can’t authenticate or query the DB during startup (the endless “mysqladmin ping” in your logs shows it’s trying but failing).
Volume corruption: Old files in /app/onlyoffice/CommunityServer/data or /app/onlyoffice/ControlPanel/data are incompatible with the new image.
.NET Core runtime crash: Missing dependencies or bad config (the “Development” mode should show more logs, but it’s not).The “no server started” is the backend not launching, so no migrations, no CRM.
My logs are stuck with this error at the end:
+ sleep 1
+ mysqladmin ping -h onlyoffice-mysql-server -P 3306 -u onlyoffice_user --password=onlyoffice_pass --silent
mysqladmin: [Warning] Using a password on the command line interface can be insecure.
Any help would be greatly appreciated. I would rather not roll back to yesterday and lose 5 hours x 3 people of work!
Edit: I got the portal loading page back but then broke it again because I cant get the community server started and connected to the database.
Edit 2: Its still down. The issue seems to be with the mysql database being 5.7 and incompatible with any new versions and for some reason I cant change any permission in the logs folders, so the process of starting the docker back up fails.
edit 3: Got the portal up and loaded with my database backup from hours earlier. But cant log in. Tried multiple different approaches, including finding the admin ID and hash and changing the password. Also tried making a new admin user. Cant login. “Invalid username or password.”
Final edit: ended up running the forgot password feature once I realized the SMTP settings were still intact. The only issue now is that the documents server has been emptied. It shows all the files, but when you try and select one, it says download failed. All documents and files linked in the CRM are dead too. @Alexandre any ideas?
Ken
