n/a

Thanks Medim i couldnt find the project id anyways

i thnk i got it 6e426d89-591b-42ce-b8dc-9ac489140a01

You got the port wrong, use the one the TCP proxy provided

pg_restore -U postgres -h tcp://switchyard.proxy.rlwy.net -p 37627 -W -F t -d railway zedtopialatest.dump

i don't know because the port doesn't show as labeled in the UI

i tried your command verbatim copied and pasted and that didn't work either. i'm open to any suggestions.

One sec, updating my postgres version so I can try it myself

if you'd like to see everything that I tried, you can see in my terminal session how I followed the tutorial and successfully ran the heroku pgdump but then when i followed the instructions to restore a pgdump on Railway it doesn't work.

I tried every method of connecting to my instance.

intead of the pghost which gives the first attempt, i then tried the database url, proxies, tcp proxies, and even deleted and redid the tcp proxy to no avail.

i changed the command and fixed the syntax from your suggestion to get pg_restore -U postgres -h switchyard.proxy.rlwy.net -p 37627 -W -F t -d railway zedtopialatest.dump

it ran but it also had restore errors....

1381708548353953792

and when i logged in and tried to run my app or login with a pre-existing user it doesn't load

most importantly. NONE OF MY USERS OR THEIR DATA WAS RESTORED.

None of the data seems to have been sucessfully migrated.

if this can't work then I can't use this platform

I hope there's a way to migrate data otherwise we're screwed.

the data doesn't restore at all. nothing shows up.

and everytime i try to click anything on my app i get a retry later error message which is infuriating.

That's because of some internal schemas and triggers that heroku uses on their dbs

https://help.heroku.com/HYME3MD5/unexpected-statements-in-the-backup-from-a-essential-x-database-is-causing-local-pg_restore-to-fail

They have a resolution there but is a very annoying fix

Some people on stackoverflow made it work by using the params --clean --no-acl --no-owner --section=pre-data --section=data

But it is weird because the heroku specific command in the guide you linked supposedly does that

i tried the link you shared from heroku and it said:

azeem@x9:~/alphazed$ pg_restore -l zedtopialatest.dump > unfiltered.list
pg_restore: [archiver] unsupported version (1.15) in file header

i just tried the new params you shared and it says:

pg_restore -U postgres -h switchyard.proxy.rlwy.net -p 37627 -W -F t -d railway zedtopialatest.dump --clean --no-acl --no-owner --section=pre-data --section=data
Password: 
pg_restore: error: could not execute query: ERROR:  cannot drop table public.users because other objects depend on it

the output is still running. would i have to delete or reset my database on railway and then try to restore again each time?

here's what it said:

1381722085658726410

i think i was able to get it to work using the information you shared

i had to make sure the version of pg dump and pg restore were all the same as my psql, and equal to or newer than the heroku version

1381738088660992180

you can use this update the readme. i see my updated data on the new site now. Thanks for helping and finding that heroku documentation

1381738741982302319

so make sure the versions are the same, dump using the tutorial and restore using the lists method with the syntax me and you built.

i got a majority of data to load now however there are still issues.

i just got Unknown primary key for table comments in model Comment.

so did this not copy over the primary keys?

when i dropped the db and tried to redo the procedure with a clean db to verify i don't see the new information populate either

none of the restores have worked after attempting to drop to restore from clean slate.

after the first attempt went thru not a single restore has worked since.

has anyone ever actually successfully restored a backup from heroku on railway in 2025?

Did you try with a fresh deployed postgres instance or did you use the same one as before?

i did a redeploy and that did not work.

Could u please try deleting that postgres instance, deploying a new one and running the restore command (the last one that worked) again?

So we can take it from there

ok

i deleted the postgres instance, and all that shows now is my github repo and a fish-volume

i added a new postgres and it automatically added a new volume, so i deleted the old fish volume of the old postgres, it's provisioning right now

clean instance, now just ran the command with the new host address, pg_restore -U postgres -h yamanote.proxy.rlwy.net -p 19052 -W -F t -d railway --use-list filtered.list --no-owner zedtopialatest.dump and it ran with no errors but no data was migrated

so i login with the ssh and try to run db:create or db:migrate since it shows there are no tables yet, thus it had nowhere to put the data i tried to restore.

Can you use a db client to connect and check to see if the railway table actually exists?

NO TABLES EXIST AND IT DOESNT EVEN CONNECT

literally the repo can't connect to the new database

do i have to delete the whole project or something? why is this platform so shitty?

No, just the postgres service, does it says that it is running fine?

i deleted the whole project already because this is bullshit

This wasn't supposed to be so hard, Railway spins up a plain postgres instance and serves it (not a custom one like Heroku)

what you say and the reality are two very different things

none of your tutorials are up to date

I wonder if we can dump it without the extra data from Heroku

the only time it successfully imported some data it omitted keys on comments etc

Heroku uses Amazon Aurora for DBs and I can assure you I exported and imported multiples dbs from that service to Railway

lol then why doesn't it work today?

Let's go from the beggining, let's start with the dump

Dump it again, use this command:

pg_dump DB_URL --no-acl --no-owner --no-privileges --clean --if-exists > dumptorailway.sql

Then, before we proceed let's check if your Railway postgres instance is running fine, what's showing in the logs?

(I updated the command above, I forgot about the --no-acl)

ctrl K doesn't show the import variables from heroku interface either.

i can't even try dumping it again yet because the new repo doesn't import variables from heroky

heroku*

typing import variables into the ctrl K box doesn't work unless you're at the variables screen. they should have just added a button this ctrl K interface is garbage. i just redid the whole thing now i have to delete the database url and replace it with railways

when i railway ssh it doesn't show the new project so i can't run a db:migrate

Don't use ssh for such things

lmfao you haven't added a way to run a migration post deploy

https://docs.railway.com/guides/pre-deploy-command

i unlinked from the old project and linked to the new one

yes i said POST deploy, not pre

you install the new code and new migrations from the repo then run them.

and then if you want to run a one off rake task etc you have to ssh as well

anyways i ssh'ed and ran rails db:migrate so we have a new clean repo and db with migrations but empty

I think you're skipping things a bit

what did i skip?

i made a whole new project and redid everything

Wasn't our focus to migrate your postgres db first?

no it was to restore my postgres db

yeah, from heroku to railway

well those two words mean different things and this is tech support bro

i used to have this job at IBM

so anyways i can't run a ruby on rails application without running rails db:migrate after the first install.

i did that. now my user table exists etc.

And the data?

Primary keys are fine?

the data we want to restore now is the question do we dump again or try the old command again?

We dump only the data.

so last time i dumped i did pg_dump -h cc6tobl4vo5pjf.cluster-czrs8kj4isg7.us-east-1.rds.amazonaws.com -d dealh3bcbmv4hg -U udajirh8klodnn -p 5432 -W -F t > zedtopialatest.dump

how do you want to modify this?

If we already have the tables and structure, we should only dump the data.

  --data-only \
  --no-owner \
  --no-privileges \
  --no-acl \
  --schema=public \ 
> public_data_only.sql```

I added the --data-only and the --schema=public

ok

and for the restore we gonna use psql instead of pg_restore, that dump returned a .sql file as you can see

  -h host-here \
  -p port-here \
  -U postgres \
  -d railway \
  -W \
  -f public_data_only.sql

fill in the host-here and port-here since you redeployed your db

i dumped using azeem@x9:~/alphazed$ pg_dump -h cc6tobl4vo5pjf.cluster-czrs8kj4isg7.us-east-1.rds.amazonaws.com -d dealh3bcbmv4hg -U udajirh8klodnn -p 5432 -W -F t > zedtopialatest.dump --data-only --schema=public Password:

did it error?

nope

it's supposed to be a .sql file lol

oh lol should i redo?

just change .dump to .sql

U also forgot the rest of the flags but ok I guess

can you just type the command correctly so i can copy it

i'm tried

tired*

azeem@x9:~/alphazed$ pg_dump cc6tobl4vo5pjf.cluster-czrs8kj4isg7.us-east-1.rds.amazonaws.com

-p 5432

--data-only

--no-owner

--no-privileges

--no-acl

--schema=public \

public_data_only.sql

pg_dump: error: too many command-line arguments (first is " ")

pg_dump: hint: Try "pg_dump --help" for more information.

edited it

I deleted the -h and user

by mistake

i did

pg_dump -h cc6tobl4vo5pjf.cluster-czrs8kj4isg7.us-east-1.rds.amazonaws.com -d dealh3bcbmv4hg -U udajirh8klodnn -p 5432 -W -F t > zedtopialatest.sql --data-only --schema=public --no-owner --no-privileges --no-acl
Password: 

it ran no errors

so now do we just restore?

Yeah, try it

how's the deployed DB on railway looking like?

it has all my tables but it's empty

so exactly a new repo with a migrate ready to run

ok

i didn't run the restore or psql yet

so to confirm, should i try the restore or the psql?

psql

i already know this syntax isnt gonna work:

psql \
  -h host-here \
  -p port-here \
  -U postgres \
  -d railway \
  -W \
  -f public_data_only.sql

and the PGhost variable is postgres.railway.internal

no it isn't

it's the tcp proxy

your platform gives me postgres.railway.internal when i copy PGHOST

i know that this syntax is wrong and we need tcpip link

the link for this new postgres is gondola.proxy.rlwy.net:40138

psql

-h gondola.proxy.rlwy.net

-p 40138

-U postgres

-d railway

-W

-f public_data_only.sql

ok i ran it

no errors

let's see if it worked or actually pulled data

the tables are still empty

we successfully ran the commands but the data isn't there

ok let's add verbose

ran the verbose no issues, but no data is showing up either.

Use a db client to look at data

also, can you open that .sql file and see if there's any insert statement?

if it ran without errors then the issue is with our dump

nah filename error

azeem@x9:~/alphazed$ psql

-h gondola.proxy.rlwy.net

-p 40138

-U postgres

-d railway

-W

-f zedtopialatest.sql

this is running

but there are errors

i'm waiting and i'll send the complete response

oh yeah true, your filename is different

i'm getting a shitton of syntax errors

got a whole bunch of

Query buffer reset (cleared).
psql:zedtopialatest.sql:10237: error: invalid command \neverything
psql:zedtopialatest.sql:10238: error: invalid command \N
Query buffer reset (cleared).
psql:zedtopialatest.sql:10240: error: invalid command \n
Query buffer reset (cleared).
psql:zedtopialatest.sql:10241: error: invalid command \n
Query buffer reset (cleared).
psql:zedtopialatest.sql:10242: error: invalid command \n
Query buffer reset (cleared).
psql:zedtopialatest.sql:10243: error: invalid command \n
Query buffer reset (cleared).
psql:zedtopialatest.sql:10245: error: invalid command \n
psql:zedtopialatest.sql:10246: error: invalid command \N
psql:zedtopialatest.sql:10247: error: invalid command \N
Query buffer reset (cleared).
psql:zedtopialatest.sql:10248: error: invalid command \n
Query buffer reset (cleared).
psql:zedtopialatest.sql:10249: error: invalid command \n
psql:zedtopialatest.sql:10250: error: invalid command \N
psql:zedtopialatest.sql:10251: error: invalid command \N
Query buffer reset (cleared).
psql:zedtopialatest.sql:10252: error: invalid command \n
psql:zedtopialatest.sql:10253: ERROR:  syntax error at or near "7856"
LINE 1: 7856 “Whosoever of you sees an evil, let him change it with ...
        ^
psql:zedtopialatest.sql:10253: ERROR:  syntax error at or near "and"
LINE 1: and if he is not able to do so, then [let him change it] wit...
        ^
psql:zedtopialatest.sql:10902: ERROR:  syntax error at or near "and"
LINE 1: and if he is not able to do so, then with his heart — and th...
        ^

also alot of ``` ^

psql:zedtopialatest.sql:2431: ERROR: syntax error at or near "base64"

LINE 1: base64,iVBORw0KGgoAAAANSUhEUgAABG0AAAIuCAIAAACKEGJoAAAAAXNSR...

^

psql:zedtopialatest.sql:2431: ERROR: syntax error at or near "width"

LINE 1: width: 50%;

^

psql:zedtopialatest.sql:2431: ERROR: syntax error at or near "">I clicked the like button because I like the feature. I clicked the like button because I like the feature. I clicked the like button because I like the feature. I clicked the like button because I like the feature. ...

^

psql:zedtopialatest.sql:2431: ERROR: syntax error at or near "">Here are the sources that were linked:Here are the sources that were linked:

and no data is in the table:

Ok, i've found some docs

let's dump this again...

lol ok

for your ease of use in giving a new command here's the old one pg_dump -h cc6tobl4vo5pjf.cluster-czrs8kj4isg7.us-east-1.rds.amazonaws.com -d dealh3bcbmv4hg -U udajirh8klodnn -p 5432 -W -F t > zedtopialatest.dump

lmk the new one

successfully ran no errors

how do you want to restore?

amazing.

should i try pg_restore or psql?

pg_restore --verbose --clean --no-acl --no-owner -h gondola.proxy.rlwy.net -p 40138 -U postgres -d railway -W latest.dump

pg restore, its a binary not a .sql file

oops

I put in the heroku creds

use the railway one

pg_restore --verbose --clean --no-acl --no-owner -h cc6tobl4vo5pjf.cluster-czrs8kj4isg7.us-east-1.rds.amazonaws.com -p 5432 -U udajirh8klodnn -d dealh3bcbmv4hg -W herokudump.dump ?

use the railway creds

oh

yea

that too

so different cred and filename

pg_restore --verbose --clean --no-acl --no-owner -h gondola.proxy.rlwy.net -p 40138 -U postgres -d railway -W herokudump.dump

how's it goin?

ok let's use psql then

psql --verbose -h gondola.proxy.rlwy.net -p 40138 -U postgres -d railway -W -f herokudump.dump

btw is the file called latest?

"latest.dump"

nah you put Herokudump.dump

sorry all lowercase

psql

-h gondola.proxy.rlwy.net

-p 40138

-U postgres

-d railway

-W

-f herokudump.dump

then run this again with herokudump.dump

ok

if it errors then we try the psql

i did already i told you

8:19

ok then try the psql one

ok gonna try

psql \
    -h gondola.proxy.rlwy.net \
    -p 40138 \
    -U postgres \
    -d railway \
    -W \
    -f herokudump.dump

forgot the verbose tag but ok.

it's giving a ton of syntax errors again like the last time

you used latest.dump here instead of herokudump.dump

did u run this command again but with herokudump.dump?

medim i ran the command i just pasted

yeah, and it errored out

I want you to run this one since when u ran it you used the wrong file

you gave me a command with the wrong file last time

yeah, my mistake

and bad syntax

just take your time and give me something correct

This is the command the heroku docs tell us to run pg_restore --verbose --clean --no-acl --no-owner -h gondola.proxy.rlwy.net -p 40138 -U postgres -d railway -W herokudump.dump

I copy pasted it

and forgot to change latest.dump to our actual dump name which is herokudump.dump

That's why it throwed an error saying "error: input file is too short (read 0, expected 5)"

Can you try it again but with the correct name file?