Hi.
I have a 600GB PostgreSQL database, and I want to set up replication. The issue is that the process of performing pg_basebackup and starting the slave database takes about 3 hours. After that, it seems the replica (slave) attempts to fetch WAL files from the primary (master), but they have already been removed.
Could this issue be related to the wal_keep_size or other WAL-related configurations?
What is the recommended value for these settings to handle a lag of 3 hours?
My Postgres version is 11.
https://postgrespro.com/community/books/introbook
https://www.postgresql.org/docs/17/index.html
It's is not possible
You can use pgpool-ll or manage this from application end like using 2 dB urls.
https://stackoverflow.com/questions/25061136/haproxy-splitting-read-write-postgresql
https://www.percona.com/blog/configure-haproxy-with-postgresql-using-built-in-pgsql-check/
Читать полностью…Can I make haproxy automatic separate read from write traffic on single port , the same port for read and write?
Читать полностью…yes, and for the read/write separation. Their best guess as to what is read/write is probably good enough to get them started. And I would take EVERYTHING that they are not sure about, and find a way to execute it.
WHERE you can check to see if it consumed an actual TXNID as a result.
Finally, I would see if there were a way to forcibly tell the connection you want a READONLY connection.
Potentially by using a read_only user.
And exercise those questionable calls to see if they error out...
The overall recommendation is to read the ACTUAL PostgreSQL Documentation.
It's Really Good.
Personally, when I started, AI was not there. I would ask it to summarize the docs for me.
Then I will focus on what I know and work with the AI to compare/contrast to Oracle (b/c I know oracle).
And that's how I would start. But I would also be doing HANDS on stuff, like installing PG.
Installing extensions. Writing Code in plpgsql (stored procedures).
I always like to take something from a different database and try migrating it over.
HTH
PGPool-II can split by analysing the queries, but I'd highly advice against doing that, especially if you use store functions/procedures.
You can also do similar with multiple databases defined in PGBouncer if your application has the concept of different read vs write connection strings.
Does anyone know any good "introduction to data modeling" courses or texts? I haven't kept up with the current state of publication in that area, and a friend asked me about it.
Читать полностью…I'm a Mongodb DBA now I want to learn the postgresql database from scratch .how I start or any link or video or online training for postgresql database please guide
Читать полностью…I realized that my question was not at all specific. Thank you for your answer! I don’t trust AI at all yet, although there is a future in this. Thank you again 🙏
Читать полностью…Is there online or offline literature on this Database Management System so that I don’t ask stupid questions here? ) Sorry please
Читать полностью…