Troubleshooting¶
Docker no space left on device¶
Docker unauthorized¶
Suppose you see an error message like this when trying to run an experiment using Docker:
ERROR: failed to solve: failed to copy: write /var/lib/docker/buildkit/content/ingest/ae8153b11f4d4f00d8b937b5de83ad657bae8a815251f89f9476de4147382577/data: no space left on device
This means too many old Docker images have accumulated on your system. This can be fixed by running the following command:
docker system prune
Database connection refused¶
Suppose you see an error message like this:
connection to server at "localhost" (::1), port 5432 failed: Connection refused
Is the server running on that host and accepting TCP/IP connections?
This means that your local Postgres database cannot be accessed. This would normally only happen if you are not using PsyNet through Docker.
If you are on a Mac, you can check the status of your database by running this command:
brew services
If you don’t see a line with postgresql, you have not installed PostgreSQL.
Follow the virtualenv installation instructions to do so.
If you do see a line with postgresql, it probably has error written next to it.
You need to get access to the logs to debug this error.
To do so, look at the File column of the brew services output,
find the value corresponding to postgresql. Print that file in your terminal using cat,
for example:
cat ~/Library/LaunchAgents/homebrew.mxcl.postgresql@14.plist
Look for a line like this:
<key>StandardErrorPath</key>
The error log path is contained underneath it, between the <string> identifiers.
View the last few lines of that file in your terminal using tail, for example:
tail /usr/local/var/log/postgresql@14.log
Have a look at the error message. One possible message is something like this:
2023-04-25 16:53:51.224 BST [28527] FATAL: lock file "postmaster.pid" already exists
2023-04-25 16:53:51.224 BST [28527] HINT: Is another postmaster (PID 716) running in data directory "/usr/local/var/postgresql@14"?
If you see this error message, try restarting your computer and trying again.
Another possible error message is this:
Reason: tried: '/usr/local/opt/icu4c/lib/libicui18n.72.dylib' (no such file)
It has proved possible in the past to fix this problem by running the following:
brew reinstall postgresql@14
brew services restart postgresql@14
where postgresql@14 should be replaced with the exact name for the Postgres service that you saw in brew services.
If that doesn’t work, try searching Google for help. If you find another solution, please share your experience here.
MISCONF Redis is configured to save RDB snapshots¶
If you you see an error beginning ‘MISCONF Redis is configured to save RDB snapshots’, and you are using MacOS, then you may be able to fix your problem by running the following command:
brew services restart redis
Postgres stops working after a Homebrew upgrade¶
If you find that Postgres stops working after upgrading via Homebrew, you might need to delete your local Postgres files and try again. This can be done as follows (these instructions are from Moncef Belyamani’s tutorial):
brew remove --force postgresql
Or if you had previously a versioned form of Postgres, for example Postgres 14:
brew remove --force postgresql@14
Delete the Postgres folders:
rm -rf /usr/local/var/postgres/
rm -rf /usr/local/var/postgresql@14/
Or if you’re on an Apple Silicon Mac:
rm -rf /opt/homebrew/var/postgres
rm -rf /opt/homebrew/var/postgresql@14
Finally you can reinstall Postgres:
brew install postgresql@14
brew services start postgresql@14
Heroku Startup Error / CLI not responding¶
If you’re experiencing issues with Heroku CLI not responding or encountering errors, you can try uninstalling and reinstalling it first. After that, enable debugging environment variables to get more detailed logs and information about what might be going wrong. This can help you diagnose and resolve the problem more effectively.
If the CLI is not responding or you’re experiencing issues, you can try uninstalling and reinstalling it.
brew uninstall heroku
rm -rf ~/.local/share/heroku ~/Library/Caches/heroku
brew install heroku/brew/heroku
After uninstalling and reinstalling the CLI, try running the Heroku command again.
If the issue persists, open your terminal and set the following environment variables to enable debugging:
export HEROKU_DEBUG=1
export HEROKU_DEBUG_HEADERS=1
export DEBUG=*
HEROKU_DEBUG=1: Enables debug logging for the Heroku CLI.
HEROKU_DEBUG_HEADERS=1: Enables debug logging for HTTP headers, useful for troubleshooting authentication or networking issues.
DEBUG=*: Enables debug logging for all modules used by the Heroku CLI.
To verify your CLI installation, use the heroku --version command:
heroku --version
If the CLI is installed correctly, you should see output similar to:
heroku/7.0.0 (darwin-x64) node-v8.0.0