[pgpool-general: 8366] pgpool failing over on bad queries

[Jon SCHEWE]

I'm finding that I have a client sending a bunch of bad queries to pgpool. After a large number of these errors pgpool appears to decide that the primary backend is in a bad state and fails over to another backend. Why is pgpool doing this and how do I prevent it? Do I need to change failover_on_backend_error to off?

I understand that the right answer is to fix the application. However I don't want my pgpool cluster to keep switching backends because someone writes some bad SQL.

The bad SQL results in the following messages showing up in my log:
2022-09-01 17:09:52.877: PostgreSQL JDBC Driver pid 2306964: LOG:  Unable to parse the query: "select * from ($1) as result" from client 192.1.192.6(51076)
2022-09-01 17:09:52.878: PostgreSQL JDBC Driver pid 2306964: LOG:  Error message from backend: DB node id: 1 message: "syntax error at or near "$1""



Jon Schewe

Principal Software Systems Technologist

C: +1 612.263.2718

O: +1 952.545.5720

jon.schewe at raytheon.com

Raytheon BBN

Raytheon Intelligence & Space

5775 Wayzata Blvd. Suite 630

St. Louis Park, MN 55416


RTX.com<https://www.rtx.com/> | LinkedIn<https://www.linkedin.com/company/raytheontechnologies> | Twitter<https://twitter.com/raytheontech> | Instagram<https://www.instagram.com/raytheontechnologies>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pgpool.net/pipermail/pgpool-general/attachments/20220901/f2cecb4f/attachment.htm>