connect ECONNREFUSED 172.22.0.2:3001
Welp - I had this working and I got an error. I noted that there was a need for a server upgrade and tried that but nothing improved. I trashed my stack and started over and this is what I'm left with. I still can't even find a sign on page from inside the server. I'm using portainer and attached the setup files (and logs from server and proxy). I can't figure out why it stopped working, but if anyone can help with getting it set up and usable, I'd appreciate it. Thanks!
111 Replies
From which version did you upgrade?
Looks like an issue with typesense. The logs were cut off, but it looks like typesense is unreachable from the immich-server. Can you confirm the typesense container is up and running?
At this point, I'm unsre what version it was. I don't have any records of the previous stack.
typesense appears to be running, (green "running" button) but there's no logs and no ipaddress noted.
No IP address seems odd. Logging is disabled in the compose file. You can try commenting it out to see the logs.
the only thing I see is logging: [blank], do I just toss a # in front or type in "enable" or something like that in the field?
If you put a # in front of logging and driver: none that should do it
You can bring the container down, then remove the typesense container and its volume then bring it back up
I brought the container down, removed it and the corresponding volume, added the "#" in front of the logging and diver: none lines and redeployed the stack. Still no ipaddress for the typesense and if I open the log it shows: "No log line matching the '' filter"
Can you pose the modified docker-compose file?
We use typesense 0.24.1 btw
ooh - I hadn't noticed that. I'll redownload the compose file from github and try it all again.
redownloaded and redeployed - everything remains the same; still no ipaddress and no log
Idk what to say. They should all have IP addresses.
Maybe delete the images locally and re pull. Not sure what else to do if you don't see any errors. If you do docker ps -a you could maybe see an exit code which could give s hint
I noticed the # TYPESENSE_ENABLED=false, do I need to uncomment and make true? (by the way, I tried that and the same thing occurred).
Forgive the ignorance, what is ps -a?
I redeployed and there was a brief moment where there was an ip address, but it only last that moment. I clicked on logs and still had the filter error.
docker ps will print out the status of the containers, -a will include containers that have stopped (a as in all i believe)
Where do I enter that command?
Terminal on the machine running portainer. There might be a way to see the exit code in prayer itself. Not sure on that. Maybe under containers or container details?
Hi, having the same problem when updating to v1.62.1.
I had the same issue yesterday but today after doing the same thing it just worked
@gaxx what do you mean by doing the same thing?
I had same error creating portainer stack, didn’t change anything in the default configuration and couldn’t run it. This morning I wake up, copied same docker compose and env files and it worked perfectly
Please try to completely bring down the stack, then pull new images and then bring the stack up again
Tried that. Removed all images still got the same. Then removed all volumes. still same. Will debug and see if I can get some more logs. Always get docker exit code 139 (SIGSEGV) from typesense
You shouldn’t remove the volume otherwise all your instance metadata will be lost
Yeah I learned that. but have snapshot to restore
Yes, I tried that also. Didn’t really help. I also should mention that this was completely fresh install with all the defaults that it comes with. Only thing is I did that all through portainer
What do you guys see in the log?
which system/architec are you guys running on?
if i run docker-compose up i get the following errors
ENOTFOUND getaddrinfo ENOTFOUND typesense and from immich_typesense: docker container exit with code 139. Running on x86 in docker in LXC container.can you remove these lines in typesense container, bring down and up again then check the log of typesense
Sorry but I won't really provide any more info (or confirm that my logs said the same thing) as this just simply worked for me today! I am using x86, virtualized debian with portainer. I only joined discord because of this issue, searching for answers. Wanted to ask myself but found this thread. I was reluctant of posting anything until I try fresh in the morning and violla, somehow it worked :)
this is the only part of the log from immich_typesense
Can you provide your docker-compose file?
using unmodified docker-compose and .env with only changed typesense api key from here https://github.com/immich-app/immich/blob/main/docs/docs/install/docker-compose.md
been running for 6 months so this is not a new instance
what is the output of your
docker ps?
Can you help posting the docker-compose of yours for the sake of completeness?
You should be using this https://github.com/immich-app/immich/releases/download/v1.62.1/docker-compose.yml instead of the file from mainare they not the same?
not always,
main is developing branch so if we change it before the release, the release file will be differentah, ok good to know. In this case they seem to be the same 🙂
Why don't we get the latest compose file from release
then take down all the container
and perform pull again
finally bring them back up
Do you mind provide screenshot for each steps?
typesense/typesense:0.24.1 "/opt/typesense-serv…" 12 minutes ago Restarting (139) 4 seconds ago immich_typesense
docker psONe more thing
ah how about we do this step by step together?
docker-compose down
then sudo systemctl restart docker
then what is the outpuf of docker ps?nothing
Good
now perform
docker-compose pulldone
now
docker-compose upsame result 😦
hmm
ok one more thing
docker-compose down
then docker images what is the output?i have these old
altran1502/immich-server
alsodocker image prune
after that can you remove the typesense image?
will do
then after that
docker-compose pull againsame result again...
immich_typesense exited with code 139
bring down the container
remove the typesense volume
then bring the container up
same thing again
Hmm
when was your last working version?
pretty sure it was v1.61.0
but was running a older docker-compose with
Why don't we try that version of typesense
working
with 0.24.0 🤔
if you use 0.24.1 does it work?
without the pinning hash
I think you have a corrupted / invalid image. I don't think it's specific to 0.24.0. I think if you deleted all the images for typesense and then re-pulled it new it would probably work.
0.24.1 breaks things again
lol patch version breaks some system... damnit Typesense
@jrasm91 @thomas FYI
I don't think it's a problem with the image. Can you delete it and re-pull it?
already done
typesense <3
How did you remove the image?
docker image rm de9eadc78acd
That was the only image for typesense?
yeah
thanks for the debugging help👏
Super strange. I think we've always used 24.1
yeah but we up the patch over the last release
if i do
docker run typesense/typesense:0.24.0 i get output but docker run typesense/typesense:0.24.1 just crashesGitHub
Typesense 0.24.1 Docker container exits with code 139 · Issue #1003...
Description I have VPS server running some services and would like to try Typesense without changing my current file system. In order to do that, I'm trying to run Typesense inside a Docker con...
could be some docker version problems
solved by updating to latest docker-engine. Thanks all!
The latest version on Debian is 24 and this is fixed in 25?
Cause I am now having the same issue after updating
docker version that is
propably
why don't you update your docker-engine and try again?
there is nothing to update
its already the latest version
immich-app docker version
Client: Docker Engine - Community
Version: 24.0.2
API version: 1.43
Go version: go1.20.4
Git commit: cb74dfc
Built: Thu May 25 21:52:17 2023
OS/Arch: linux/amd64
Context: default
interesting
maybe revert back to 0.24.0 for typesense
that is after a docker compose down / docker compose pull / docker compose up -d
okay and you have errors?
when I try to view the logs in the typesense container it says it cannot retrieve logs
becasue it is disable by default
if your server can start up fine then you are good to go
ok .... its working....
I tried multiple times over the last 10 minutes and kept getting 502
just tried it again and its working...
if you have lots of assets, when the stack starts, typesense will take sometimes to index/load data into memory and the server will have to wait until that process is completed before finishing the boostraping sequence
i have the same issue, but i'll give it some time
(i have waited 30 mins and it fixed the job)
hearing some of the "tried again this morning and it just worked", I tried removing the stack and starting again/repulling the image and it's still the same thing.
I'm using portainer on a NAS, so for a terminal, would I SSH into portainer?
The host running portainer most likely
I assume you mounted the docker socket when running portainer as a docker container itself.
I have no idea what you just said (I feel like Ted Lasso in the first episode). I have an Asustor and there's apps for both docker and portainer (the former must be installed for the latter to work)...so I'm guessing all the mounting and such was done automatically.
still no ip address for typesense container (after repulling)
Can you run
docker ps -a in a terminal on the machine that is running portainer?I don't see any terminal app on the Asustor NAS, which is running the Docker/Portainer combo. I tried to use powershell to ssh and could get in, but when running the command, I get:
|Got permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock: Get "http://%2Fvar%2Frun%2Fdocker.sock/v1.24/containers/json?all=1": dial unix /var/run/docker.sock: connect: permission denied
trying again, loggin in as root
got it
typesense restarting (139)
Can you put the output here of
docker version?
And can you verify if using the typesense image 0.24.0 instead of 0.24.1 makes any difference?I tried using 0.24.0, with the same results
Can you update docker please?
another dumb question...how can I do that when I downloaded the docker engine from the Asus store (I have the most recent version available there)?
You might want to uninstall it from there and install it directly from the docker website instead
I wish I knew how to do that - I've tried searching for how to's and such and I come up with nothing. If you have any suggestions, I'm all ears.
When you use 0.24.0 did you pull the image again?
Try restarting the system
The linked issue specifically mentions updating docker as a fix
Ok...so it sounds like I need to figure out how to update docker.
I still can't find anything on command line installations on an Asustor (there must be something, I just haven't found it yet).
I can do a manual install through the App Station, but it seems to only accept .opkg
I'll see if I can figure out how to bundle the necessary elements into an appropriate file.
Yeah, it looks like the docker engine app is just old. Not sure if there is a way to request an update to it.
The alternative would be to enable ssh access, login as an admin and then use that to install it instead.
But that leads to a complete reset of all accounts
While using
docker compose down -v and up again fixed it, it leads to a complete reset
All my photos are gone too 😭
But they're there in the upload location
Can someone explain that-v flag remove the volume which is the database
I still have all the photos. Will the live photos be as they were before?
yes they will be auto link when you reupload
How do you identify live photos in a folder?
Yup that was a blunder. Learnt a huge lesson today 😭😂
they usually come in pair, so one
.heic and one .mov with the same nameI thought that too. Just wanted to confirm. Thanks 👍
Is it the same with jpegs? Like one .jpeg file and one .mov file?
I suppose
It looks like it is maintained by Asus, I' have a message into them to update their Docker version. In the meantime, I can ssh in - I need to verify my system architecture and grab the correct binary)
As a side bar, I attempted to see if I could build the APK for Asustor, and I don't think I've found any straight answers. Hopefully I can get them to update it (or get lucky and find some answers on the APK front).
Where are you seeing the engine app age? I ask because I did get an answer from Asus and they said they update it every few weeks. I'm hoping to show them what you are seeing and try to figure out a solution.