BUG #16605: PostgreSQL recovery startup process waiting and blocking to application queries

From: PG Bug reporting form <noreply(at)postgresql(dot)org>
To: pgsql-bugs(at)lists(dot)postgresql(dot)org
Cc: tushar11(dot)takate(at)gmail(dot)com
Subject: BUG #16605: PostgreSQL recovery startup process waiting and blocking to application queries
Date: 2020-09-02 10:58:49
Message-ID: 16605-174a554a8e2f6385@postgresql.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

The following bug has been logged on the website:

Bug reference: 16605
Logged by: Tushar Takate
Email address: tushar11(dot)takate(at)gmail(dot)com
PostgreSQL version: 10.5
Operating system: OEL 6.9
Description:

Hi Team ,

Problem/bug/issues statement : - PostgreSQL recovery startup process is
blocking the application queries .

Env Details -

PostgtreSQL community version - 10.5
OS - OEL 6.9
Replication type - Streaming - async

DB parameters at replica side -

max_standby_streaming_delay = -1
max_standby_archive_delay = -1
hot_standby_feedback = off

Startup process by ps -ef

postgres 24489 24487 0 2019 ? 1-20:23:35 postgres:
startup process recovering 00000006000021D4000000ED waiting

Startup process from pg_stat_activity [Here its not showing process is in
waiting state ]

postgres=# select * from pg_stat_activity where pid='24489';
-[ RECORD 1 ]----+------------------------------
datid |
datname |
pid | 24489
usesysid |
usename |
application_name |
client_addr |
client_hostname |
client_port |
backend_start | 2019-06-17 22:45:33.679547+00
xact_start |
query_start |
state_change |
wait_event_type |
wait_event |
state |
backend_xid |
backend_xmin |
query |
backend_type | startup

Startup process blocking all the application queries more than 4 hours .

2020-08-29 09:50:07 UTC [4596]: [3-1]
user=postgres,db=pgprd_production,app=psql,client=[local] LOG: process 4596
still waiting for AccessShareLock on relation 17280 of database 16408 after
1000.076 ms at character 15
2020-08-29 09:50:07 UTC [4596]: [4-1]
user=postgres,db=pgprd_production,app=psql,client=[local] DETAIL: Process
holding the lock: 24489. Wait queue: 25091, 27676, 27186, 28443, 24754,
27109, 27679, 26314, 28682, 24756, 25348, 26938, 30990, 29405, 28770, 29408,
25351, 27625, 26091, 26470, 29170, 30756, 28683, 28589, 27195, 29336, 25694,
30637, 25191, 25346, 29660, 27872, 6563, 6625, 26491, 29492, 30443, 28767,
27742, 28877, 25629, 5941, 26311, 27993, 6951, 25026, 30581, 27746, 30527,
30757, 6564, 6566, 29582, 27188, 30085, 25021, 27041, 25352, 27187, 27042,
29173, 26028, 25073, 29649, 26793, 25194, 28590, 7666, 11080, 26808, 26798,
28528, 27499, 24948, 29694, 13580, 6444, 30684, 6565, 30532, 4596.
2020-08-29 09:50:07 UTC [4596]: [5-1]
user=postgres,db=pgprd_production,app=psql,client=[local] STATEMENT: select
* from mytable limit 1;

All are the pid which are in wait queue are executing the select queries on
same table that is (mytable) .

Can you please help us to understand the issues and possible
solution/workaround/fix for it .

For temp fix we have restarted the DB service ( Before that we have tried to
kill the process by -15 flag which not worked )

Thanks,
Tushar Takate .

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Peter Eisentraut 2020-09-02 12:06:18 Re: BUG #15285: Query used index over field with ICU collation in some cases wrongly return 0 rows
Previous Message Matthias van de Meent 2020-09-02 10:43:56 Re: BUG #16586: deduplicate_items=true can be configured for numeric indexes