Metadata Storage Access


According to this document, most tables in the metadata storage are only accessed by coordinator, overlord and middleManager processes. It also mentions the “config” table but doesn’t provide much information about the processes accessing that table. Do historical processes need to access metadata storage?

My cluster is currently running on Docker Swarm and a common environment file is shared among all processes. An update to that file causes all processes to restart and I’m about to update some DB settings inside that file. I’m trying to decide whether to leave the DB settings in that file or move them elsewhere.

Thanks in advance!


Has anybody tried using connection pooling for metadata storage db connection? If so, could you share the config with me?

I’m using postgresql for storing the metadata(on aws rds) and I observe a slowness when getting the segment list. I’m thinking adding some dbcp conn pool settings such as :

but when I do that, coordinator starts to fail bc somehow conn jdbc url gets corrupted (it seems to be an empty string according to the exception)

thanks in advance