-
Notifications
You must be signed in to change notification settings - Fork 2.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Enable non-transactional migrations #2939
Comments
We can add this is a query parameter to |
@0x777 Is it possible to add this as a parameter or arg within the up and down migration files? as an alternate method. Sometimes we do want transactions, and sometimes we don't. The default can be with transactions, and with an arg they can be turned off. And then we continue to use |
@0x777 Ah I see a miscommunication on my end - the request was to not have transactions when doing migrations specifically, not really queries. |
@avimoondra Thanks for bringing this up. Could you clarify the title of the issue, please? Maybe something like "Enable non-transactional migrations". Without the ability to |
I'm looking for this too. Migrating from Django where non-atomic migrations are standard. |
This would be a much appreciate feature. We have this in Ruby on Rails which allows us to do |
This is also needed for any DDL actions for timescaledb for creating continuous aggregates etc because it uses two separate transactions internally. For now, a workaround is to manually apply the But it'd be really nice to have a CLI flag like |
Hi, is there any news on this topic? Concurrent index creations and deletions would be very nice to have! Otherwise, any updates on the linked tables would be locked during the entire duration of the migration (shameless plug of an article I've written on this topic). The downside is that concurrent index management would take more time to finish (two parses of the table are needed instead of one, and PG would wait for any transaction to finish between these steps). It can be relatively scary, even if harmless. |
I definitely would love to see this added. The more our database grows, the more we need to have indexes added concurrently which cannot be done through migrations. |
For DB safety reasons, sometimes it is necessary to add an index concurrently. With the
CONCURRENTLY
option:But this parameter cannot be used in migrations because Hasura runs the migration in a transaction when applying. But concurrent index creation cannot be wrapped in a transaction block:
There don't seem to be any workarounds or escape hatches in this case, beyond not tracking the migration at all and running directly on the db instance.
Some solutions might be to include an option in the apply api, or encoding not wrapping a transaction in the yml for the migration file itself.
Original request in Discord: https://discordapp.com/channels/407792526867693568/535727661167673364/626175223724769295
Postgres documentation: https://www.postgresql.org/docs/9.1/sql-createindex.html
The text was updated successfully, but these errors were encountered: