Re: 回复:Re: PANIC: wrong buffer passed to visibilitymap_clear

Поиск
Список
Период
Сортировка
От Tomas Vondra
Тема Re: 回复:Re: PANIC: wrong buffer passed to visibilitymap_clear
Дата
Msg-id 19330739-8460-a110-c826-8d5a44b100a9@enterprisedb.com
обсуждение исходный текст
Ответ на 回复:Re: PANIC: wrong buffer passed to visibilitymap_clear  ("王伟(学弈)" <rogers.ww@alibaba-inc.com>)
Список pgsql-hackers
On 7/22/22 14:17, 王伟(学弈) wrote:
> On 7/22/22 18:06, Tomas Vondra wrote:
>> Which PG14 version / commit is this, exactly? What sysbench parameters
>> did you use, how likely is hitting the issue?
> PG_VERSION is '14beta2'.
> The head commit id is 'e1c1c30f635390b6a3ae4993e8cac213a33e6e3f'.

Why not current REL_14_STABLE? 14beta2 is pretty old, and while I
haven't checked, perhaps this was already fixed since then.

> I have run these sysbench commands for  couple of days, but only two times to hit the issue.
> These sysbench commands are:
> prepare:
>
sysbench --tables=10 --table-size=1000000 --db-ps-mode=auto --pgsql-password=* --time=72000 --db-driver=pgsql --pgsql-port=*
>
--threads=50 --thread-init-timeout=3000 --report-interval=5 --pgsql-user=* --pgsql-host=* --pgsql-db=* --events=0 --pgsql-ignore-errors="PX000,58M01"
> ./src/lua/oltp_insert.lua prepare
> parallel execution: 
>
sysbench --tables=10 --table-size=1000000 --db-ps-mode=auto --pgsql-password=* --time=72000 --db-driver=pgsql --pgsql-port=*
> --threads=50 --thread-init-timeout=3000 --report-interval=5 --pgsql-user=* --pgsql-host=*
> --pgsql-db=* --events=0 --pgsql-ignore-errors="PX000,58M01"
> ./src/lua/oltp_insert.lua run
>
sysbench --tables=10 --table-size=1000000 --db-ps-mode=auto --pgsql-password=* --time=72000 --db-driver=pgsql --pgsql-port=*
>
--threads=50 --thread-init-timeout=3000 --report-interval=5 --pgsql-user=* --pgsql-host=*--pgsql-db=* --events=0 --pgsql-ignore-errors="PX000,58M01"
> ./src/lua/oltp_delete.lua run
>
sysbench --tables=10 --table-size=1000000 --db-ps-mode=auto --pgsql-password=* --time=72000 --db-driver=pgsql --pgsql-port=*
> --threads=50 --thread-init-timeout=3000 --report-interval=5 --pgsql-user=* --pgsql-host=*
> --pgsql-db=* --events=0 --pgsql-ignore-errors="PX000,58M01" ./src/lua/oltp_read_write.lua run
>
sysbench --tables=10 --table-size=1000000 --db-ps-mode=auto --pgsql-password=* --time=72000 --db-driver=pgsql --pgsql-port=*
> --threads=50 --thread-init-timeout=3000 --report-interval=5 --pgsql-user=* --pgsql-host=*
> --pgsql-db=* --events=0 --pgsql-ignore-errors="PX000,58M01" ./src/lua/oltp_update_index.lua run
> 

Thanks. Not sure I'll be able to do such long sysbench runs, though. Can
you try reproducing this with current REL_14_STABLE? I wonder if dumping
the WAL (using pg_waldump) might tell us more about what happened.


regards

-- 
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company



В списке pgsql-hackers по дате отправления:

Предыдущее
От: Cary Huang
Дата:
Сообщение: Re: Add last failed connection error message to pg_stat_wal_receiver
Следующее
От: Zheng Li
Дата:
Сообщение: Re: Support logical replication of DDLs