Thread: Auto vacuum not reclaiming the space and disk utilization spike to 100%

Hi Team,

I am running  Postgres v12.2 and in our testing we are running performance test for 72 hrs with load. During the run it is working fine till first 16hrs after that disk utilization getting spike and next 2-4 hrs it is  reaching to 100%.

During the run we are deleting 2 million records from one of the partition table’s older  partition every 20 second. As we are deleting such records gradually, auto vacuum process is executing accordingly. But I can see few partition which have clean all the records but its size is still getting in MBs.

nftdb=# select count(*) from XXX_sequence_pp18;

count

-------

     0

I am suspecting due to this disk utilization is getting higher in my environment.

Can you please help to tune such issue

Thanks & Regards,

Ishan Joshi