public inbox for [email protected]help / color / mirror / Atom feed
Routing of FDW Queries to Secondary databases 3+ messages / 2 participants [nested] [flat]
* Routing of FDW Queries to Secondary databases @ 2026-04-17 19:11 Adam Blomeke <[email protected]> 2026-04-22 05:36 ` Re: Routing of FDW Queries to Secondary databases Tatsuo Ishii <[email protected]> 0 siblings, 1 reply; 3+ messages in thread From: Adam Blomeke @ 2026-04-17 19:11 UTC (permalink / raw) To: [email protected] Hi, I understand that FDW queries don't get routed to the secondary databases since they occur within FETCH blocks. Is there any planned work to change this. I have one database I'm about to work on clusterizing, but this limitation makes pgpool impractical as the read side is only accessed via FDWs in all but a very small set of circumstances. Again, I realize the existing limitation; I'm just curious if there's any work on the horizon to add that capability. If not, I will likely need to explore other alternatives for one of the clusters I'm working on. Cheers, Adam ^ permalink raw reply [nested|flat] 3+ messages in thread
* Re: Routing of FDW Queries to Secondary databases 2026-04-17 19:11 Routing of FDW Queries to Secondary databases Adam Blomeke <[email protected]> @ 2026-04-22 05:36 ` Tatsuo Ishii <[email protected]> 2026-04-23 17:55 ` Re: Routing of FDW Queries to Secondary databases Adam Blomeke <[email protected]> 0 siblings, 1 reply; 3+ messages in thread From: Tatsuo Ishii @ 2026-04-22 05:36 UTC (permalink / raw) To: [email protected]; +Cc: [email protected] > Hi, > > I understand that FDW queries don't get routed to the secondary databases > since they occur within FETCH blocks. Yes, currently Pgpool-II does not load balance DECLARE CURSOR/FETCH (it's documented). The difficulty is, once DECLARE CURSOR is routed to particular server, subsequent FETCH must be routed to the server too. Pgpool-II has to track this. > Is there any planned work to change > this. As far as I know I haven't noticed a request to implement load balancing for DECLARE CURSOR/FETCH up to now. > I have one database I'm about to work on clusterizing, but this > limitation makes pgpool impractical as the read side is only accessed via > FDWs in all but a very small set of circumstances. > > Again, I realize the existing limitation; I'm just curious if there's any > work on the horizon to add that capability. If not, I will likely need to > explore other alternatives for one of the clusters I'm working on. If there's a strong demand, developers will discuss on implementing the feature for the next major release of Pgpool-II (planned to be released by the end of this year). Regards, -- Tatsuo Ishii SRA OSS K.K. English: http://www.sraoss.co.jp/index_en/ Japanese:http://www.sraoss.co.jp ^ permalink raw reply [nested|flat] 3+ messages in thread
* Re: Routing of FDW Queries to Secondary databases 2026-04-17 19:11 Routing of FDW Queries to Secondary databases Adam Blomeke <[email protected]> 2026-04-22 05:36 ` Re: Routing of FDW Queries to Secondary databases Tatsuo Ishii <[email protected]> @ 2026-04-23 17:55 ` Adam Blomeke <[email protected]> 0 siblings, 0 replies; 3+ messages in thread From: Adam Blomeke @ 2026-04-23 17:55 UTC (permalink / raw) To: Tatsuo Ishii <[email protected]>; +Cc: [email protected] A lot of our connections operate over foreign data wrappers. We have very large datasets and we try to segment our analysts off from the actual data sources. I'd like to be able to balance my read-only FDW queries onto a secondary database. I've got 2 PGPool instances already, and this would be beneficial on one of them along with this third one. I realize I'm only one guy, but having this would be really useful in my implementation. Cheers, Adam On Tue, Apr 21, 2026, 22:36 Tatsuo Ishii <[email protected]> wrote: > > Hi, > > > > I understand that FDW queries don't get routed to the secondary databases > > since they occur within FETCH blocks. > > Yes, currently Pgpool-II does not load balance DECLARE CURSOR/FETCH > (it's documented). The difficulty is, once DECLARE CURSOR is routed > to particular server, subsequent FETCH must be routed to the server > too. Pgpool-II has to track this. > > > Is there any planned work to change > > this. > > As far as I know I haven't noticed a request to implement load > balancing for DECLARE CURSOR/FETCH up to now. > > > I have one database I'm about to work on clusterizing, but this > > limitation makes pgpool impractical as the read side is only accessed via > > FDWs in all but a very small set of circumstances. > > > > Again, I realize the existing limitation; I'm just curious if there's any > > work on the horizon to add that capability. If not, I will likely need to > > explore other alternatives for one of the clusters I'm working on. > > If there's a strong demand, developers will discuss on implementing > the feature for the next major release of Pgpool-II (planned to be > released by the end of this year). > > Regards, > -- > Tatsuo Ishii > SRA OSS K.K. > English: http://www.sraoss.co.jp/index_en/ > Japanese:http://www.sraoss.co.jp > ^ permalink raw reply [nested|flat] 3+ messages in thread
end of thread, other threads:[~2026-04-23 17:55 UTC | newest] Thread overview: 3+ messages (download: mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2026-04-17 19:11 Routing of FDW Queries to Secondary databases Adam Blomeke <[email protected]> 2026-04-22 05:36 ` Tatsuo Ishii <[email protected]> 2026-04-23 17:55 ` Adam Blomeke <[email protected]>
This inbox is served by agora; see mirroring instructions for how to clone and mirror all data and code used for this inbox