I can’t begin to describe how upset I am right this minute. I have an app in production that I finished all my requirements today. It’s been in production for weeks and with active user data. I came home from lunch to find I was unable to login to my app in dev with any user account then told the role “user” doesn’t exist. I was done! Months of work. Production is working but I can’t get any data to work in dev. BD is now gone, even backups aren’t working. I have sales demo’s for this app tomorrow. What do i do?
Are you sure it’s gone? Give more info.
check the database tab see if there is any database there, it could just be a configuration issue. in the meantime open a ticket and see if the replit team can help @kody-replit
Called with Eric yesterday and he’s back in a working state 
3 Likes
Excellent work yet again!
Yes Kody helped save the day, again; and is one of the redeeming qualities of working with Replit!! Thank You Kody!
2 Likes
Kody needs a pay rise!
1 Like
Hell yeah man, good luck on your demo!
Thanks, just finished the demo and went great! Now on to getting it cleaned up for distribution!
1 Like
I had this!! Twice! And my database content is premium! Most recent request is STILL being ignored by support after a month! Your data is not lost, and i see Kodi has helped, but for others when this happens to you, and it WILL if you’re doing serious stuff, then here’s what to do:
- Your db string will have changed. The whole thing: host and pwd. Data is not gone, your app is just looking in the wrong place. No one has been able to explain to me why this happens. But with me it definitely happened when creating forks of my app. AI is of no help here.
- So you need to keep a backup of your db string on your PC, else you likely have it in an old forked repl
- You can update this string under secrets, BUT neon/commands shows the new bad string, and you canNOT change that. And it will replace your secret every restart. So you’re screwed! And this is my support request that has been ignored for the last month.
- But agent will help you import from the original string to the new db. It’ll cost you a day of effort, frustration, and about $50 in agent costs. But it’ll work. A huge issue in importing is the created_at in every record. Omg, I’ll leave you guys to spend your own $50 getting past that.time conversion. I have removed every single time strong from my db so as never to lose these 2 days again.
- In the interim, you can hard code the original string and get your app fully working. Just do not deploy or even git with that string so visible.
- Now things get worse when the db dissapears completely from replit. And going to neon has only ONE button…to create a new db. This will literally make your balls drop the first time you see it. And no, you cannot go to neon directly cos our replit strings do not work outside of replit.
- So create the new db, and then you have the above issue so solve as above.
- For everyone else, implement a database backup button asap!! There is another post on getting pg_dump to work. Get it working!!! Every other database backup approach takes around 4 hours of the agent screwing around, and over $10 of costs. I have had to do this at least four times this year. Pg_dump can be run from shell once you have tools v17 working. It’s worth every second of effort to do this work around until replit enable the latest PG tools in their standard libraries system.
I have spent hundreds of dollars on replit. At least 70% of that has been managing the rogue agent, and this lost db issue. Ie being charged whilst replit sorts itself out. I don’t mind that, it’s early vibe coding days. I do mind that a margin has been made off me at this time.