mirror of
https://github.com/AppFlowy-IO/AppFlowy.git
synced 2026-03-24 12:56:59 +00:00
[GH-ISSUE #7969] AppFlowy Cloud Self-Hosting: Persistent "Incorrect Password" After Exhaustive Troubleshooting #3476
Labels
No labels
2024
2025
2026
acct mgmt
AI
automation
bug
calendar
ci
CJK
cloud
code-block
collaboration
copy-paste
database
data migration
data sync
deploy
desktop
develop
develop
documentation
duplicate
editor
editor-plugin
emoji
export
files
flutter-only
follow-up
formula
good first issue for devs
good first issue for experienced devs
grid
hacktoberfest
HACKTOBERFEST-ACCEPTED
help wanted
i18n
icons
images
importer
improvements
infra
install
integrations
IR
kanban board
login
look and joy
mentorship
mobile
mobile
needs design
new feature
new feature
non-coding
notes
notifications
onboarding
organization
P0+
permission
platform-linux
platform-mac
platform-windows
plugins
program
pull-request
Q1 25
Q1 26
Q2 24
Q2 25
Q3 24
Q3 25
Q4 24
Q4 25
react
regression
rust
rust
Rust-only
Rust-only
Rust-starter
Rust-starter
self-hosted
shortcuts
side panel
slash-menu
sync v2
table
tablet
task
tauri
templates
tests
themes
translation
v0.5.6
v0.5.8
v0.5.9
v0.6.0
v0.6.1
v0.6.4
v0.6.7
v0.6.8
v0.7.1
v0.7.4
v0.7.4
v0.7.5
v0.7.6
v0.7.7
v0.7.8
v0.8.0
v0.8.4
v0.8.5
v0.8.9
web
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
AppFlowy-IO/AppFlowy#3476
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @AristotleSays on GitHub (May 22, 2025).
Original GitHub issue: https://github.com/AppFlowy-IO/AppFlowy/issues/7969
Problem Description:
Despite successfully deploying AppFlowy Cloud via Docker Compose on Proxmox LXC, resolving multiple critical setup issues (storage, CPU compatibility, network, and service startup errors), I am unable to log in to the AppFlowy web app or desktop client. The system consistently returns "Your password is incorrect, please try again," even after confirming the admin user exists in the database with the correct password.
Environment Information:
motivationpve/data) on a 4TB NVMe SSD.192.168.3.105latest(fromAppFlowy-IO/AppFlowy-CloudGitHub repository).Network Configuration:
FioS Router --> pfSense firewall (192.168.2.1 WiFi, 192.168.3.1 Lab) --> Proxmox cluster (spark, motivation) --> motivation node --> appflowy-ct (LXC)LABinterface for192.168.3.105to allowICMPandUDP(Port 53), andTCP/UDPfor general internet access.nameserver 192.168.3.1,nameserver 8.8.8.8,nameserver 1.1.1.1in/etc/resolv.conf.Chronological Troubleshooting & Resolutions:
Initial Problem:
no such logical volume pve/data (500)when creating LXC onsparknode.sparknode'sstorage.cfghadlvmthin: local-lvmdefined, but no underlying LVM-Thin pool existed.lvmthin: local-lvmentry fromspark'sstorage.cfg. Created LXC using ZFS storage (rpool-data) onspark.Second Problem:
Fatal glibc error: CPU does not support x86-64-v2inminiocontainer logs.sparknode (Mac Pro 3,1) CPU lackedSSE4.2instruction set required by modernglibcbuilds in Docker images.vzdumpbackup ofappflowy-ctonspark. Transferred backup tomotivation'slocalstorage. Addedlvmthin: local-lvmdefinition tomotivation'sstorage.cfg. Restoredappflowy-ctbackup tolocal-lvmonmotivation. All core Docker services started (Upandhealthy).Third Problem (Current): Persistent "Your password is incorrect, please try again" on login (both web app and desktop client).
Troubleshooting Steps for "Incorrect Password" (Chronological):
Initial
.envConfiguration:APPFLOWY_BASE_URL=http://192.168.3.105:3060GOTRUE_ADMIN_EMAIL=your_admin_email@example.com(Replaced the actual value for submission here)GOTRUE_ADMIN_PASSWORD=[YOUR_SIMPLE_PASSWORD_HERE](Replaced the actual value for submission here)GOTRUE_JWT_SECRET=[YOUR_GOTRUE_JWT_SECRET_HERE](original was specific string)GOTRUE_DISABLE_SIGNUP=trueGOTRUE_LOG_LEVEL=debug(though debug output did not appear in logs)..envlikePOSTGRES_PASSWORD=[YOUR_POSTGRES_PASSWORD](originalpassword),AWS_SECRET=[YOUR_MINIO_PASSWORD](originalminioadmin), etc., were set.GoTrue
Admin user already exists: Initially,gotruelogs showed "Admin user already exists. Skipping..." preventing password update.sudo docker compose down -vto wipe all Docker volumes (including PostgreSQL data), forcing a fresh database state for user creation.gotruelogs then showed "Creating admin user for gotrue..." without "skipping".Password Characters: Attempted with
GOTRUE_ADMIN_PASSWORD='my password with spaces'(quoted) and latersimplepassword123(alphanumeric only).Web App Behavior & Email Verification: Web app login gave "Request Failed with Status Code 500" and asked for email verification.
SMTP Configuration Errors:
gotruelogs showedFATAL: assigning GOTRUE_SMTP_PORT to Port: converting '' to type int.docker-compose.ymlwas explicitly passing${GOTRUE_SMTP_PORT}(which expanded to empty string from.env) togotrue.GOTRUE_SMTP_HOST,GOTRUE_SMTP_PORT,GOTRUE_SMTP_USER,GOTRUE_SMTP_PASS,GOTRUE_SMTP_ADMIN_EMAILlines fromgotrue'senvironment:section indocker-compose.yml.appflowy_cloudlogs then showed the sameFATALerror regardingAPPFLOWY_MAILER_SMTP_PORT.docker-compose.ymlwas explicitly passing${APPFLOWY_MAILER_SMTP_PORT}(empty string) toappflowy_cloud.APPFLOWY_MAILER_SMTP_HOST,APPFLOWY_MAILER_SMTP_PORT,APPFLOWY_MAILER_SMTP_USERNAME,APPFLOWY_MAILER_SMTP_EMAIL,APPFLOWY_MAILER_SMTP_PASSWORD,APPFLOWY_MAILER_SMTP_TLS_KINDlines fromappflowy_cloud'senvironment:section indocker-compose.yml.Upandhealthy. Web app no longer asks for email verification.Manual Password Update in DB (Final Attempt):
Action: Used
python3 bcryptto generate a new hash forsimplepassword123.Action: Directly updated the
encrypted_passwordforyour_admin_email@example.comin theauth.userstable in PostgreSQL. Confirmed update viaSELECTquery. ExampleSELECToutput:Outcome: Still "Your password is incorrect, please try again" on both web app and desktop client.
Current Status:
All AppFlowy Cloud services are running and appear healthy in
docker ps. The database user (your_admin_email@example.com) exists with a confirmed, manually updated password hash. There are no critical errors ingotrueorappflowy_cloudlogs indicating startup failure or password verification issues. The web app is accessible, but login consistently fails with "Incorrect Password."Request for Assistance:
Given the exhaustive troubleshooting and verification that the user exists and the password is correct in the database, the persistent "incorrect password" error is highly unusual. It points to a very subtle issue during the password verification step that is not being logged as an error, or a client-side interaction problem.
Any insights or suggestions for further debugging steps would be greatly appreciated. Thank you.
@khorshuheng commented on GitHub (May 23, 2025):
Thanks for the detailed debugging steps. I am assuming, that by web, you are trying to log in to the admin panel (/console), instead of the web version of the Appflowy (/), because the web version doesn't support password login at the moment (there is no field to even provide the password).
If that is the case, I would try debugging this using the admin panel first. Try checking the browser dev console, and see which API endpoint returns the error. It is likely going to be the same reason why the login failed for the desktop app.
@AristotleSays commented on GitHub (May 24, 2025):
I am self-hosting the server. Can you provide more details? I have exhausted my troubleshooting skills. Please give me clear instructions (step-by-step, if possible) as I am new to AppFlowy application. I didn't even know this application existed until 2 days.
@khorshuheng commented on GitHub (May 24, 2025):
What I meant is, try logging into admin panel using username and password. The admin panel should be on /console, if setup correctly. And if you are not able to do so, the browser console should tell you more on the actual error.
@raylinger commented on GitHub (May 26, 2025):
i have the exact same issue and did exactly the same as OP.
macos Sequoia 15.5; Docker Compose, pull latest version, everything works, every service shows healthy and up, set a user + pw in the .env file (restarted) and it does not work with provided credentials.
I deleted everything multiple times and also tried it multiple times with the admin@example.com template that's already set in the .env file.
In Web there is no possibility to sign in without a magic link and with the /console i just get a random string, which stretches over the whole screen (can't copy/paste it). In the App it shows the incorrect password message, no matter how many users i create, how simple / hard the pw is, with or without special characters.
@khorshuheng commented on GitHub (May 26, 2025):
If you get the long random string, it is probably issue with verifying the token, which you might see error either in the AppFlowy cloud server, or on the browser dev console.
Also, just to eliminate some of the more common issues - make sure you pull the latest image by docker compose pull. In addition, the password of the admin user does not get overwritten when you change the env variable.
@raylinger commented on GitHub (May 29, 2025):
hello,
thank you for the feedback.
At this point i am positive, that i have tried everything. The things you mentioned, are common sense.
Either i miss something in the .env that needs to be changed/configured, or it simply is broken.
I could do a no-cut setup video and upload if it is not believable... :)
Update: when i type some user that does not exist, i get a corresponding message, invalid user credentials.
If i type the correct user and password, i get the long string. there is no information in the dev console to read except generic error 500
@J-P-Meth commented on GitHub (Jun 4, 2025):
Hello,
i have exactly the same problem, with a similar setup.
When i try to log into
/console, with the default login (admin@example.com) I receive this message as response in the Firefox console:When I change the
ADMIN_FRONTEND_APPFLOWY_CLOUD_URLvariable in the.envfile, to the local IP address of my AppFlowy instance, then I receive this message:(Filled the string with
*)i don't know what is the problem here.
@khorshuheng commented on GitHub (Jun 4, 2025):
Chances are, AppFlowy cloud service failed to reach some service. For example, if you have modified the name of the gotrue service but didn't modify the env accordingly. There should be more details in the logs in the AppFlowy cloud service.
@J-P-Meth commented on GitHub (Jun 6, 2025):
I can't find any errors in the docker logs itself.
What do you mean with 'modified the name of the gotrue service'?
When i do a
docker ps -ai get this:Is it correct, that there is no Ports for the gotrue container?
@khorshuheng commented on GitHub (Jun 6, 2025):
Hmm...ok, on the first glance it does look correct. Could you share your .env variable ( without the sensitive information)?
@khorshuheng commented on GitHub (Jun 6, 2025):
I am also assuming that, there isn't any other proxy in front?
@J-P-Meth commented on GitHub (Jun 6, 2025):
Thanks alot for the fast response.
I don't use any other proxy in front.
My
.envfile looks like this.@khorshuheng commented on GitHub (Jun 6, 2025):
Try changing the admin AppFlowy cloud service back to the original value (appflowy_cloud) again. One possibility that I can think of, is that, at the time when you try to login, the container is somehow not ready yet.
Basically the error message means that the admin frontend service is not able to send a http request to appflowy_cloud:8000 - which shouldnt be the case if the container is up and running.
And I am also assuming that, you are not running docker compose in host network mode?
@J-P-Meth commented on GitHub (Jun 6, 2025):
The login works now :)
Thank you very much.
I don't know why it was not working before (with the value appflowy_cloud)
I don't running docker compose in host network mode.
The sync between Linux App and the ios App is not working properly.
I need to log out and log in again, that the client is syncing.
But that is a other Problem.
Thanks a lot again.
@annieappflowy commented on GitHub (Jun 13, 2025):
@J-P-Meth , I'm closing this issue for now. We'll be rolling out a new syncing protocol soon, which will resolve many of the past issues and offer more reliable performance. We look forward to getting your feedback!