Re: Hash Right join and seq scan
On Fri, 5 Jul 2024 at 12:50, James Pang wrote: >we have a daily vacuumdb and analyze job, generally speaking it's done in > seconds, sometimes it suddenly running more than tens of minutes with same > bind variable values and huge temp space got used and at that time, explain > show "Hash
Re: How to solve my slow disk i/o throughput during index scan
On 7/4/24 22:23, FREYBURGER Simon (SNCF VOYAGEURS / DIRECTION GENERALE TGV / DM RMP YIELD MANAGEMENT) wrote: Hello, Thank you, splitting in “OR” query definitely enables bitmap heap scans, and thus parallelized read to disk 😃! I though did not understand your second point, what is parallel app
Re: Hash Right join and seq scan
we have a daily vacuumdb and analyze job, generally speaking it's done in seconds, sometimes it suddenly running more than tens of minutes with same bind variable values and huge temp space got used and at that time, explain show "Hash Anti join, Hash Right join" with seq scan two tables. Tom
RE: How to solve my slow disk i/o throughput during index scan
2024-07-04
Thread
FREYBURGER Simon (SNCF VOYAGEURS / DIRECTION GENERALE TGV / DM RMP YIELD MANAGEMENT)
Hello, Thank you, splitting in “OR” query definitely enables bitmap heap scans, and thus parallelized read to disk 😃 ! I though did not understand your second point, what is parallel append, and how to enable it ? Simon F. Interne De : Andrei Lepikhov Envoyé : jeudi 4 juillet 2024 16:37 À
Re: How to solve my slow disk i/o throughput during index scan
On 4/7/2024 20:25, FREYBURGER Simon (SNCF VOYAGEURS / DIRECTION GENERALE TGV / DM RMP YIELD MANAGEMENT) wrote: *My question is : what can I change to get a better index reading speed ?* What I already tried : * Setting random_page_cost to prohibitive value (1000) to force a bitmap hea
