PostgreSQL

25 February 2026 07:54

Where did the original post go?

PostgreSQL

23 February 2026 13:02

Maybe there is an update that isn't applying on the replica or it hasn't done a checkpoint etc.?

PostgreSQL

23 February 2026 12:52

And the question re the subscriber meant: did you check that there's really a replica connected? Sometimes a firewall suddenly appears out of nowhere, you know...

PostgreSQL

23 February 2026 12:51

Then the setting doesn't apply anyway.

PostgreSQL

23 February 2026 12:11

"off" may disturb queries on the replicas. Anyway, if your _slot_ is blocking VACUUM, did you check there's an actual subscriber on that slot?

PostgreSQL

23 February 2026 11:59

Physical replication slots don't keep xmin (only actual replication connections with hot_standby_feedback=on do).

PostgreSQL

22 February 2026 15:29

All feasible IMHO, although you'll not enjoy the inheritance approach.

PostgreSQL

21 February 2026 06:21

Lucky he didn't run into the problems with 18.2. 😅 We had lists of journals disappearing because titles and descriptions were subject to a substring clause to limit length printed and the values contained non-ASCII characters, and I think some emails not sent on similar grounds. Had to install a snapshot version with it fixed

PostgreSQL

18 February 2026 18:54

Does anybody have pdf for “PostgreSQL 16 Administration Cookbook”.. Thanks in Advance

PostgreSQL

18 February 2026 15:23

I don’t think it’ll be that bad

Just…the thought of rolling out timeline tracking and auditing is painful

PostgreSQL

18 February 2026 15:21

Probably because it is very rigid algorithmically.

PostgreSQL

18 February 2026 15:13

Why strange performance?😅

PostgreSQL

18 February 2026 14:32

Is that the same as Oracle?🤔
I heard oracle or something of the sort has it too

PostgreSQL

18 February 2026 14:07

DB2 has at least some of the requested features.

The main advantage of implementing what's in the standard would be that ORMs could use the features. Which - come to think of it - sounds terrifying 😅

PostgreSQL

18 February 2026 09:58

According to commitfest, no activity in that field since 2021. Which is kind of sad. On the plus side it's something you can earn some merits with a good patch, I guess 😏

PostgreSQL

25 February 2026 07:42

I would immediately invest that 10.8k and keep looping. It is like printing your own money for free.

PostgreSQL

23 February 2026 12:57

yes, replica is active and slot is active true, same vlan theres no firewall or something, actually theres 2 replicas one of running well, update xmin but one of is stuck keeps same xmin , and I facing with this problem on another clusters too

PostgreSQL

23 February 2026 12:51

but if theres no activity why it keeps xmin, aint is a bug

PostgreSQL

23 February 2026 12:49

what you mean actual subscriber, its physical active rep slot, without any activity

PostgreSQL

23 February 2026 12:07

thank you, I need set to off this config right

PostgreSQL

23 February 2026 11:57

I think postgresql has a bug for physical replication slots, sometimes replication slot keeps xmin that inhibits autovacuum clean dead tuples in lider even theres no transactions in replicas

PostgreSQL

22 February 2026 14:17

How would you implement polymorphic datatype with Postgres? Would you just use JSONB or Class Table Inheritance or just nullable fields?

PostgreSQL

21 February 2026 02:55

Linux 7.0 Shows Significant PostgreSQL Performance Gains On AMD EPYC Review - Phoronix
https://www.phoronix.com/review/linux-70-amd-epyc-turin/2

PostgreSQL

18 February 2026 15:24

So much so that I have deferred the feature for so long

But it can’t wait anymore, I guess. Especially now that there doesn’t seem to be any movement in that area

PostgreSQL

18 February 2026 15:22

Most definitely not free, if it’s coming from Oracle

PostgreSQL

18 February 2026 15:16

Not an Oracle guy, but AFAIK they have this "flashback" feature. No idea if that's a) standard compliant b) free or comes with costs (guess so).

PostgreSQL

18 February 2026 14:46

Just with a strange performance and even stranger modifications!

PostgreSQL

18 February 2026 14:31

I have a feature I’ve managed to defer for almost 2 years now because I was waiting for it to happen

I think WITH SYSTEM TIME AS OF is easier if it comes baked in

That and, using a managed database, I can’t install the periods extension so bummer

PostgreSQL

18 February 2026 13:56

I don't think it's sad.

SQL as a standard is already overbloated with app-level features.

And temporal/bitemporal is mostly app-level (at leas as of now). Nothing special is of urgent need in the engine, just another set of fields and index search.

So, it seems better to wait while there are would be some implementations to look at them, look what they need, look for common features — and then implement something in the server core...

PostgreSQL

18 February 2026 07:44

Does anyone know about the current state of bitemporal features in PostgreSQL (in built, not with extensions)?

Last I remember, work was being done to include that