139
votes

I'm developing an app on Heroku with a Postgresql backend. Periodically, I get this error message when trying to access the database, both from the CLI and from loading a page on the server:

psql: FATAL: remaining connection slots are reserved for non-replication superuser connections

Anyone seen this before or please help point me in the right direction?

5
Having the same problem. I read somewhere that the Heroku support had "detected some issues" on some servers, and they recommended the user in question to provision a new Basic database and migrate to it using pgbackups. My problem now is that the app is so new that I didn't yet create a backup, and I get the same error trying to create one of course :D - André Laszlo
For the record, I was able to use heroku pgbackups command to create a backup despite this error. - markshiz
To reproduce this issue you can create large amount of terminals. .bat script in Windows for this: for /l %%x in (1, 1, 100) do ( start psql ) where 100 is desired number of backends. - koxt
I had the same problem. I wasn't able to transfer the data to a new database using pg:backups copy, pg:backups capture, by connecting to it from pgAdmin on my own computer, or any other way I could imagine. Even pg:killall didn't help. An hour later, I tried again and connections were 50-50, so after a few attempts I had a successful pg:backups copy and my app is back in the air. This was... not a fun day. If you googled this, drink a cup of water. - Aur Saraf
I'm now having this again. How lucky to have documented my previous experience... Edit: this time restarting was enough. - Aur Saraf

5 Answers

66
votes

You either need to increase the max_connections configuration setting or (probably better) use connection pooling to route a large number of user requests through a smaller connection pool.

https://wiki.postgresql.org/wiki/Number_Of_Database_Connections

16
votes

This exception happened when I forgot to close the connections

9
votes

See Heroku “psql: FATAL: remaining connection slots are reserved for non-replication superuser connections”:

Heroku sometimes has a problem with database load balancing.

André Laszlo, markshiz and me all reported dealing with that in comments on the question.

To save you the support call, here's the response I got from Heroku Support for a similar issue:

Hello,

One of the limitations of the hobby tier databases is unannounced maintenance. Many hobby databases run on a single shared server, and we will occasionally need to restart that server for hardware maintenance purposes, or migrate databases to another server for load balancing. When that happens, you'll see an error in your logs or have problems connecting. If the server is restarting, it might take 15 minutes or more for the database to come back online.

Most apps that maintain a connection pool (like ActiveRecord in Rails) can just open a new connection to the database. However, in some cases an app won't be able to reconnect. If that happens, you can heroku restart your app to bring it back online.

This is one of the reasons we recommend against running hobby databases for critical production applications. Standard and Premium databases include notifications for downtime events, and are much more performant and stable in general. You can use pg:copy to migrate to a standard or premium plan.

If this continues, you can try provisioning a new database (on a different server) with heroku addons:add, then use pg:copy to move the data. Keep in mind that hobby tier rules apply to the $9 basic plan as well as the free database.

Thanks, Bradley

6
votes

I actually tried to implement connection pooling on the django end using:

https://github.com/gmcguire/django-db-pool

but I still received this error, despite lowering the number of connections available to below the standard development DB quota of 20 open connections.

There is an article here about how to move your postgresql database to the free/cheap tier of Amazon RDS. This would allow you to set max_connections higher. This will also allow you to pool connections at the database level using PGBouncer.

https://www.lewagon.com/blog/how-to-migrate-heroku-postgres-database-to-amazon-rds

UPDATE:

Heroku responded to my open ticket and stated that my database was improperly load balanced in their network. They said that improvements to their system should prevent similar problems in the future. Nonetheless, support manually relocated my database and performance is noticeably improved.

0
votes

To reproduce same issue in Linux:

for i in {1..300}; do
     PGPASSWORD=MY_PASSWORD gnome-terminal -e  $'/usr/bin/psql -h \'127.0.0.1\' -p 5432 -U MY_USERNAME' 
done

In a dotnet client you can read:

  System.InvalidOperationException: An exception has been raised that is likely due to a transient failure.
   ---> Npgsql.PostgresException (0x80004005): 53300: sorry, too many clients already