The following bug has been logged on the website:
Bug reference: 15965
Logged by: Shiwangini Shishulkar
Email address: shiwangini.shishulkar@webgility.com
PostgreSQL version: 11.4
Operating system: CentOS Linux release 7.6.1810 (Core)
Description:
It was kind of very strange behavior, I have observed so far. we have 69GB
PROD database and during it's backup , I can't perform any DDL and DML on
postgres database. Only 'select' statements I can run. The database takes
~45minutes for backup. Below statement I'm running for backup:
pg_dump -U postgres -d wg -f wg.sql
Looks like pg_dump locks all tables of database during backup window. Any
other way/command to take backup? Also, I have seen that postmaster CPU
utilization reaches 100% when pg_dump backup running on. Any way to avoid
this.
I have also set priority of backup script by 'nice - 18 backup.sh' , but
when backup starts postmaster gets kicked off and cause 100% CPU
utilization. Any way to avoid this?
Please get back to me in case of any query. Please try to look into this on
priority since this is happening on PROD.