v2 login not found on fresh install (v4-rc.2)
On a fresh install of v4-rc.2 (using Docker) running on my local machine for testing, I always get the error
{"code":5,"message":"Not Found"}, I presume this is the new v2 UI. Is one of my options misconfigured?
I initially started testing the latest stable release, but wanted to check out the new login form and to be able to build my own in the future. Do I need to run a reverse proxy for /ui/v2/login running with the zitadel/typescript repo?
Here is my docker compose file, based on the original provided docker compose mentioned in the docs.
41 Replies
hey @Vaungo welcome and good morning, yes its failing for me too, I will check with my team. Please run any othet older versions if its a fresh deployment(only for testing)
🎉 Looks like you just helped out another community member! Thanks for being so helpful <@403882453653127168>! You're now one step closer to leveling up—keep up the amazing peer support! 🚀
+1, happening for me too with all v4 release candidates
hey @Light I have raised this internally, I will get back to you all soon 🙂
@Vaungo @Ben did you find a solution to this? I am still experiencing this, even on
v4-rc3 :/hey @Light yes for me to, v4-rc3 is broken and identfied already, i fix is on the way
Awesome!
sounds like that issue may be because you have the wrong path set for the v2 login or that v2 isn't exposed on the right path
hey @Ozzzkar what was the path you set for the V2 login
When I first installed Zitadel, login V2 was working. But then after I pressed a button to reset some settings in the web console, login V2 stopped working and instead login V1 showed up. Then I had to configure the login V2 base URL manually in the setting and then it started working again after setting it to this
https://my-zitadel-domain.example.com/ui/v2/login
It sounds to me like OP's issue is that the traffic going to the V2 URL doesn't end up at the V2 container but rather at the V1 container
Which is the same container as the other Zitadel components are running in
hey @Ozzzkar can you pls share your doker file?
You mean values.yaml for the Helm chart? I installed using Kubernetes + Helm
can I see your values.yaml for the helm chart, just curous to see the setup 😄
i'm using the latest version of your helm chart https://github.com/zitadel/zitadel-charts/tree/zitadel-9.0.0-rc.2/charts/zitadel
and my values are roughly these:
hey @Ozzzkar after discussing with my colluages I think you were correct about actions v1, but its not an issue but a known limitation that won't be fixed since Actions v1 are being replaced by Actions v2
same for me with the released 4.0.0 (with postgresql 17)
docker logs don't tell much, everything looks fine:
zitadel | time="2025-08-01T16:02:11Z" level=info msg="server is listening on [::]:8080" caller="/home/runner/work/zitadel/zitadel/cmd/start/start.go:677"
when I try to navigate to http://localhost:8080/ui/console -> I get a redirect to /ui/v2/login/login?authRequest=V2_... with
it's a fresh installationI find it hard to believe that the Zitadel devs wouldn't run into the same issue themselves. So it must be a simple misconfiguration from our side? 🤔
have you enabled the ingress for zitadel-login? (this is the kubernetes resource name for login v2)
I am trying to run this locally with docker (compose), no k8s for now
I want to keep it as simple as possible for initial research of the product, but I'm having hard time even running it
I copied all the configs from https://zitadel.com/docs/self-hosting/deploy/compose 1:1
somehow it started, so I was able to create an org and project inside.
then I restarted the docker - everything broke down, traefik throws up with a bunch of errors:
login and zitadel show no changes in the logs.
the whole setup is so fragile.
the moment I try remove traefik from the setup - zitadel starts loosing db creds or something else.
is the general suggestion to stick with v3 or what?
is there any solution?
https://github.com/zitadel/zitadel/issues/10379
I made a ticket
I didn't see any similar tickets there, as they try to reroute everyone to discord
How can it be that you know this issue, have not fixed it and still released 4.0.0 as a stable release?
There is a workaround to do the setup with v3 and then upgrade to v4 but still…
the path is part of the docker compose file so that looks right
I wonder if running
docker compose up -d would make any difference. Maybe something got stale when Docker restartedhey guys
@Ozzzkar @borales @Vaungo
the issue to this error of npot found was very simple actually, I only got the response from my engineers just now.
can you try it with all files in a clean directory and with fresh data?
docker compose down -v
mkdir /tmp/clean
cd /tmp/clean
and then try running all the wget/docker commands form here
https://zitadel.com/docs/self-hosting/deploy/compose
docker thinks it should mount not yet existing directories (not files) and it creates the directories. This breaks future attempts of docker compose up.
That is why you kept on getting this error
@Rajat thanks, it let me in.
Although when I add a JWT key for my API app - they are not saved (the list of keys remains empty).
I attached a few log lines from zitadel. Any idea what could be wrong?
still something's wrong
even without using jwt and only code - I get gatewate timeouts
I just ran into the exact same issue on a fresh v4.0.0 installation directly on Linux (no docker). I thought I was getting insane, it is quite annoying
Hi @Rajat any updates on the fix?
hey @Light pls share your docker file
Of course @Rajat , here is my docker compose and the env we use for zitadel:
hey @Light it's fixed on rc4 and on latest , can you pls try, I just checked with my engineers
Yess I will test it right now :)
works!
https://localhost/ui/console?login_hint=zitadel-admin@zitadel.localhost
try this link ^^ once your containers are running
@Rajat I just wiped everything (images, volumes, etc.) and started it up again. Unfortunately I run into the same "Not found" error again. I can't use your URL as my traefik does the routing to zitadel.
Could you maybe send me your compose and env so I can compare and see if I have sth missing?
mkdir /tmp/clean
cd /tmp/clean
and then the commands from the docs again
Not found is v2 missing. That’s what I get if o boot v4 on sec hosted without a v2 login
OK .. maybe this is the issue with me -- i've done so many
docker compose up --force-redeploy my head is spinning.Hey, have you found a way to make it work ? I am having this exact issue and its driving me insane.