Is it right that Postgres 14.5 coming this Thursday will be the first minor 14 update without any fixes related to [CREATE INDEX | REINDEX] CONCURRENTLY?
(looking at gitlab.com/postgres/postg)
Conversation
Not sure of the specifics offhand, but I believe that the CIC bugs discovered recently were found because Yandex have been using amcheck + heapallindexed on their fleet. Mostly not a 14 thing.
1
4
Right but as I understand the last fix (in 14.4) was specifically related to optimization in 14 and which unfortunately was reverted due to it -- teach RC&CIC not to hold xmin horizon and not to block dead tuple deletion
Previous fixes where not only for 14 though, right..
2
1
That's true, of course. Though even that was found in the same way (by Andrey Borodin and/or his colleagues).