PostgreSQL really is the world's best database, but isn't perfect.https://medium.com/@rbranson/10-things-i-hate-about-postgresql-20dbab8c2791 …
-
-
Replying to @rbranson
Have you experienced the same magnitude of wraparound issues with 9.6+? After that PG doesn't again scan unchanged parts of a table, even during an anti-wraparound vacuum. Combined with the less insane autovacuum defaults in v12 it has reduced (but not removed) the pain IME.
1 reply 0 retweets 4 likes -
Replying to @AndresFreundTec @rbranson
FWIW, I've seen very busy setups using syncrep, with some geo distribution (mostly US / EU). And others fail badly. Depending on the workload the locking isn't a problem (it mostly matters if there's row-level contention).
1 reply 0 retweets 2 likes -
Replying to @AndresFreundTec @rbranson
Syncrep can be enabled/disabled on a per-transaction basis, which allows it to be used in more situations. Often a high % of tx don't have that high durability requirements. The latency impact is paid once per transaction, and can often be amortized across parallel commits.
1 reply 1 retweet 2 likes
Your description of full page writes is off btw, significantly overstating the overhead. Not required for every modification, only if page has not been modified since the last checkpoint. Enabling checksums can address a significant part of the "corruption propagation" concern.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.