a year ago
for some reason when both the ui and i try to connect to my postgres instance, we seem to be getting failures. the error message i am seeing is
psql: error: connection to server at "monorail.proxy.rlwy.net" (000.000.000.000), port 00000 failed: server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
wondering how to approach this / debug it or maybe there is something i am doing wrong. my start command is
docker-entrypoint.sh -c 'shared_buffers=1GB' -c 'temp_buffers=32MB' -c 'max_connections=500' -c 'work_mem=256MB' -c 'maintenance_work_mem=256MB' -c 'max_wal_size=1GB' -c 'listen_addresses=*'
my docker container is https://hub.docker.com/r/gibsfinance/assets-postgres
the only other difference is that i am using
/var/lib/postgresql/data/pgdata13
for my PGDATA value. i did this change in an effort to get around this weird error that i was seeing where the db was complaining that the database was created by postgres v16, when i built and everything is using v13
0 Replies
what is more is i don't even see any connection break in the logs. they are just chilling, waiting to accept connections
a year ago
can you check to see if the database is using the legacy or V2 runtime?
a year ago
please also provide the deploy logs for the database, and a screenshot of the tcp proxy
a year ago
can I get that screenshot please?
oh, wait, is the proxy forwarding to a different port? / do i need to fix the pg port on 5432?
a year ago
beat me to it
a year ago
though whenever you can, definitely connect to the database via the private network so you can avoid the egress fees
ah, interesting. ok. so the "private" network is available from outside of railway?
a year ago
nope, wouldn't be very private then
a year ago
that's why I said whenever you can, because you aren't going to be able to use the private network to access the database from your local computer, but your app that runs in the same environment will be able to
a year ago
let me know if you run into any more difficulties or have any more questions
actually, i am having some difficulty for some reason with accessing the db from other services.
when i read, it says that it is unable to find the row i am looking for, which is correct, the row does not exist
the migrations seem to have worked, but during data collection/seeding i hit this issues of being unable to aquire connection
i am also having difficulty for some reason downloading submodules that i access during seeding
a year ago
can you share the specfic error?
' at Client_PG.acquireConnection (/usr/src/app/node_modules/knex/lib/client.js:311:13)\n' +
' at Runner.ensureConnection (/usr/src/app/node_modules/knex/lib/execution/runner.js:305:46)\n' +
' at Runner.run (/usr/src/app/node_modules/knex/lib/execution/runner.js:30:30)\n' +
' at Target.then (/usr/src/app/node_modules/knex/lib/builder-interface-augmenter.js:24:43)\n' +
' at process.processTicksAndRejections (node:internal/process/task_queues:95:5)'
a year ago
thats only 5 log lines, please use this -
the other error is the no such file or directory, scandir
line which i also find odd because it should have downloaded during the docker build
a year ago
Dockerfile or nixpacks?
should be dockerfile { "$schema": "https://railway.app/railway.schema.json", "build": { "builder": "DOCKERFILE", "dockerfilePath": "Dockerfile.worker" }, "deploy": { "runtime": "LEGACY", "numReplicas": 1, "cronSchedule": "0 0 * * * *", "startCommand": "npm run collect", "sleepApplication": false, "restartPolicyType": "ON_FAILURE", "restartPolicyMaxRetries": 10 } }
a year ago
i dont even think 0 0 * * * *
is a valid cron expression
a year ago
would you even want to be running a worker on a cron schedule? a worker is usually a long lived task
oh, it's just what i called it. it does one task for a couple minutes then exits
a year ago
oh okay cool
a year ago
looks like whatever frontend library railway is using can parse it, i would use the more standard 0 * * * *
though
a year ago
side question, are you seeing these characters in your logs on railway too? c-200-erc-20[24D[7B[10A[2K⠼ phux
a year ago
right but do they show up in the logs window on railway?
oh, no they show up as the character. but the log line is updated so a lot of them show
a year ago
can i see a screenshot of your logs from railway?
a year ago
for the connection issue, try changing LEGACY
to V2
a year ago
thanks
ok, so i think the reason i am getting the submodules error is because it does not seem to have the submodules before the dockerfile is run
a year ago
i unfortunately wouldnt know anything about submodules
a year ago
whats stopping you from making the necessary modifications to your Dockerfile so that it works properly?
it's just a wierd pathway. one that i did not expect. i have to initialize a git repository, then add the origin from github, then pull all the submodules
but it seems like railway does not have them when it starts the docker build
a year ago
right, the .git folder is not pulled
am i charged for building large images? 2.45GB
is what my local is showing
a year ago
railway does not currently charge for the builder or the registry storage
ok, here's a question. i am reading files from the submodules i mentioned above
a year ago
you gotta turn off that spinner thing lol
i am seeing these fetch failures where it should be going to that path under usr
fetch failure 'pls369' -> '/usr/src/app/submodules/pulsechain-assets/blockchain/pulsechain/assets/0xB64ef51C888972c908CFacf59B47C1AfBC0Ab8aC/logo.png'
i am wondering if there are permissions that i need to be concerned about
a year ago
"fetch failure" is no too helpfull, do you have a more verbose error?
a year ago
does the file exist at that location?
is this anything to be concerned about? when i remove the active ones the ui eventually shows the top one as active
a year ago
yeah I've been seeing that a lot
10 months ago
hey @gibsfinance the team just pushed a fix for this, can you let me know if you are able to remove those extra services?
10 months ago
awsome!