I just upgraded to 1.8.4 and now cannot connect on localhost:2283
Hi, I'm new to Immich and pretty new to Docker. I'm running on Windows 10 Pro and after updating to 1.8.4 from 1.8.3, everything seems to start and remain started, but I cannot connect to the website. I'm trying to load an old iPhoto library in as an external library. It has mostly been working well. I tried scanning all 60k photos at once which did not work without frequent crashing so I have been adding one year at a time. I'm also new to discord so forgive me if I do something dumb (or several things).
47 Replies
Here's my docker-compose.yml
What error do you get? Can you make sure all the containers are running, and check their logs?
yes. i just did a docker-compose up without the -d so i should have all of it. one sec. thanks!
here are the logs.
containers seem to be running.
this is what i see when i try to connect now.
I think your typesense container is having a bad time, but I don't see any messages from it specifically in the logs
Can you grab the logs for just typesense?
sure!
typesense logs
https://github.com/immich-app/immich/issues/4714 and https://github.com/immich-app/immich/issues/4780 seem potentially related, but I'm not quite sure what's going on. @Alex @jrasm91 thoughts?
Nope.
Possibly related to the embedding auto-reindexing that I think @sogan added?
Idk if this is helpful, but to be a little more clear. I suspect that various processes "found" all 60k images when I first connected my external library, but when things started crashing. I pulled out most of those (coming out of iPhoto they are in folders by year) and reran things with just one or two years. Then, I would add a year or two at a time back when it finished processing things. Everything did seem to be working fine like that. I have tagged hundreds of faces, etc.
What is that?
#4700?
Is this the kind of thing that I would be better off starting over and just adding my photo folders one at a time from the start instead of everything all at once? I haven't really invested THAT much time at this point.
Is this helpful at all? It's a list of my images right now.
I am not sure yet, I am trying to identify where the issue is
Can you bring down the containers, then remove the typesense volume and bring the stack back up again?
ok, thanks! i'm VERY new to docker. what is the command for removing the typesense volume? i don't want to goof it up.
or is that the same as deleting the image? EDIT: i see now it's not the same, but in my list of volumes, it's not obvious to me which is the typesense volume.
these are my volumes.
It would be
immich-app_tsdata
so docker volume rm immich-app_tsdatathanks! i wil give it a shot.
no luck. here's the log.
It is indexing
wait a bit
here's an update. does it typically not start serving up web pages until this finishes? will there be a very obvious log entry that it's done?
The server will say listening in port 3001 or something and that doesn't happen until typesense is up and the webpage doesn't work until the server is up
looks like it is done
okay. well, i will just leave it alone for tonight. i just searched the logs for 3001 and there were no relevant entries. thanks.
what does the web say when you accessing it?
Removing the typesense volume (and upgrading from 0.24.1 to 0.25.1) fixed this for me
I had to disable Reverse Geocoding for it to work. It did work before on version 1.82.1, after upgrading to 1.84.0 it runs into some memory leak
Right now it says this..
This page isn’t working
localhost didn’t send any data.
ERR_EMPTY_RESPONSE
The network tab under F12 in Chrome shows two rows for localhost. One has a status of "(canceled)" and the other shows "(failed)".
super dumb question. can i just do a normal docker pull typesense/typesense:latest or do i have to do something special?
Nothing special as far as I know, latest probably not a good thing but to specify an exact version
ah, okay. thanks! il will do 0.25.1 since it seems that helped a few people.
it doesn't look like the typesense update to 0.25.1 helped me. is there any way for me to check the logs for the http reqeusts? i can't tell if they are even hitting immich.
The 0.25.1 changelog includes a lot of bug fixes for embeddings, so this might be something fixed upstream https://github.com/typesense/typesense/releases/tag/v0.25.1
Are you still getting the same error logs in Typesense?
here is my newest typesense log.
Try this:
1. Bring down the containers with
docker compose -f docker-compose.yml down
2. Delete the Typesense volume with docker volume rm tsdata (if this doesn't work, you might need to search with docker volume ls to find the name of the tsdata volume and use that)
3. Deploy with new containers with docker compose -f docker-compose.yml up --force-recreateguys, i'm an idiot. in my defense, i said i was new to all of this. sooo, i wanted to make sure things weren't weird with my docker/network stuff so i pulled httpd just to test a simple connection. i did not realize that the value on the right side of the colon(:) when picking ports is what the container is listening on. i thought those were http/https values. so once i realized that mistake, i went back and checked my docker-compose.yml and sure enough i had changed that value for immich-proxy to: 2283:8081 because I thought I was going to have a conflict with another service i had running. that immediately fixed this bigger problem. really dumb on my part. i still have a problem but will post below.
but NOW, i'm getting this error from the Immich interface.
connect ECONNREFUSED 172.19.0.8:3001 (ECONNREFUSED)
Stacktrace
Error: connect ECONNREFUSED 172.19.0.8:3001
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1595:16)
Can you post your docker-compose file please
here you go. thanks!
what is your
docker ps?here it is.
everything looks ok
can you access the web again?
actually, yes, it's working now. that might have just been the delay from starting up. i think i'm good now. so sorry for all the trouble and thanks for the help. this looks like an amazing app!
No problem, cheer!