a lot of session wait on lock relation
Hi, PGv 15.10, many session coming in doing "insert into tablexxx values" in parallel, this is a range partition tables, total 12 partitions plus one default partition, only three btree indexes, no others constraints ,no foreign key. but we see hundreds of session waiting on "lock:relation", why inserts into partition table cause "relation lock" ? Thanks, James
Re: a lot of session wait on lock relation
Hi James, Could share the related test setup details? Would like to look into this. Thanks, Chetan On Thu, 15 May 2025 at 09:27, James Pang wrote: > Hi, >PGv 15.10, many session coming in doing "insert into tablexxx > values" in parallel, this is a range partition tables, total 12 > partitions plus one default partition, only three btree indexes, no others > constraints ,no foreign key. but we see hundreds of session waiting on > "lock:relation", why inserts into partition table cause "relation lock" ? > > Thanks, > > James > -- -- Regards, Chetan +353899475147 +919665562626
Re: a lot of session wait on lock relation
On Thu, 2025-05-15 at 16:27 +0800, James Pang wrote: > PGv 15.10, many session coming in doing "insert into tablexxx values" in > parallel, > this is a range partition tables, total 12 partitions plus one default > partition, > only three btree indexes, no others constraints ,no foreign key. but we see > hundreds > of session waiting on "lock:relation", why inserts into partition table cause > "relation lock" ? Something else does; use the pg_blocking_pids() function with the process ID of a blocked backend to find out who is holding the lock. If you have hundreds of sessions, you are allowing to many connections. Yours, Laurenz Albe
Re: a lot of session wait on lock relation
Laurenz Albe writes: > On Thu, 2025-05-15 at 16:27 +0800, James Pang wrote: >> why inserts into partition table cause "relation lock" ? > Something else does; use the pg_blocking_pids() function with the process ID > of > a blocked backend to find out who is holding the lock. More specifically: the inserts are only trying to get a shared lock. If they are blocked, it's because some other operation is already holding an exclusive lock on the table and is not letting go. Look for uncommitted DDL changes. More details about that at [1]. regards, tom lane [1] https://www.postgresql.org/docs/current/explicit-locking.html#LOCKING-TABLES
Re: a lot of session wait on lock relation
thanks, we are checking partition maintain jobs ,that hold access exclusive lock. Tom Lane 於 2025年5月15日週四 下午9:24寫道： > Laurenz Albe writes: > > On Thu, 2025-05-15 at 16:27 +0800, James Pang wrote: > >> why inserts into partition table cause "relation lock" ? > > > Something else does; use the pg_blocking_pids() function with the > process ID of > > a blocked backend to find out who is holding the lock. > > More specifically: the inserts are only trying to get a shared lock. > If they are blocked, it's because some other operation is already > holding an exclusive lock on the table and is not letting go. > Look for uncommitted DDL changes. > > More details about that at [1]. > > regards, tom lane > > [1] > https://www.postgresql.org/docs/current/explicit-locking.html#LOCKING-TABLES > > >
Re: a lot of session wait on lock relation
Thanks, we are checking the partition maintain job , we have 12 partitions , each week one partition there is a default partition attached with this table and huge rows in default partition too , default partition has primary key that include partition key (time range based) too. partition job detach/drop old partition and add/attach new partition each week. when add/attach new partition , with default partition to verify that it contains no records which should be located in the partition being attached, it will full scan or use partition key(part of primary key) scan instead ? primary key (columa, partitionkey). Thanks, James James Pang 於 2025年5月15日週四 下午9:32寫道： > thanks, we are checking partition maintain jobs ,that hold access > exclusive lock. > > Tom Lane 於 2025年5月15日週四 下午9:24寫道： > >> Laurenz Albe writes: >> > On Thu, 2025-05-15 at 16:27 +0800, James Pang wrote: >> >> why inserts into partition table cause "relation lock" ? >> >> > Something else does; use the pg_blocking_pids() function with the >> process ID of >> > a blocked backend to find out who is holding the lock. >> >> More specifically: the inserts are only trying to get a shared lock. >> If they are blocked, it's because some other operation is already >> holding an exclusive lock on the table and is not letting go. >> Look for uncommitted DDL changes. >> >> More details about that at [1]. >> >> regards, tom lane >> >> [1] >> https://www.postgresql.org/docs/current/explicit-locking.html#LOCKING-TABLES >> >> >>
