Re: Make NUM_XLOGINSERT_LOCKS configurable
| От | Jim Nasby |
|---|---|
| Тема | Re: Make NUM_XLOGINSERT_LOCKS configurable |
| Дата | |
| Msg-id | 5ce30017-1d82-4a4f-8678-ce936a154e73@gmail.com обсуждение исходный текст |
| Ответ на | Re: Make NUM_XLOGINSERT_LOCKS configurable (Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>) |
| Список | pgsql-hackers |
On 1/12/24 12:32 AM, Bharath Rupireddy wrote: > Test case: > ./pgbench --initialize --scale=100 --username=ubuntu postgres > ./pgbench --progress=10 --client=64 --time=300 --builtin=tpcb-like > --username=ubuntu postgres > > Setup: > ./configure --prefix=$PWD/inst/ CFLAGS="-ggdb3 -O3" > install.log && > make -j 8 install > install.log 2>&1 & > > shared_buffers = '8GB' > max_wal_size = '32GB' > track_wal_io_timing = on > > Stats measured: > I've used the attached patch to measure WAL Insert Lock Acquire Time > (wal_insert_lock_acquire_time) and WAL Wait for In-progress Inserts > to Finish Time (wal_wait_for_insert_to_finish_time). Unfortunately this leaves the question of how frequently is WaitXLogInsertionsToFinish() being called and by whom. One possibility here is that wal_buffers is too small so backends are constantly having to write WAL data to free up buffers. -- Jim Nasby, Data Architect, Austin TX
В списке pgsql-hackers по дате отправления: