Gpcopy transfer size larger than size of table

Products

• VMware Tanzu Greenplum

Issue/Introduction

The reported size of data transferred in gpcopy  is larger than size of the tables that were copied.

gpcopy logs:

20221221:15:11:09 gpcopy:gpadmin:XEMDUB07DBXP180:027640-[INFO]:-[datalake] Finished copying database "Progress: (1/1) DBs, (1/1) tables done"
20221221:15:11:10 gpcopy:gpadmin:XEMDUB07DBXP180:027640-[INFO]:-------------------------------------------------
20221221:15:11:10 gpcopy:gpadmin:XEMDUB07DBXP180:027640-[INFO]:-Total elapsed time: 8h11m9.747886529s
20221221:15:11:10 gpcopy:gpadmin:XEMDUB07DBXP180:027640-[INFO]:-Total transferred data 528.8GB, transfer rate 65GB/h
20221221:15:11:10 gpcopy:gpadmin:XEMDUB07DBXP180:027640-[INFO]:-Copied 1 databases
20221221:15:11:10 gpcopy:gpadmin:XEMDUB07DBXP180:027640-[INFO]:-        Database datalake: successfully copied 1 tables, skipped 0 tables, failed 0 tables

Table size: 
 

select
   relname,
   pg_size_pretty(pg_table_size(relid)) as table_size
   where
   schemaname = 'schema'
 and relname = '<table name> '
 order by relname asc;

 relname | table_size 
--------------------+------------ 
mopar_ant_forecast | 58 GB

Environment

Product Version: 6.19

Resolution

The gpcopy report shows the size of the data that is transferred. The size of the table is what is on disk. If the table is stored using compression, than the actual size of data in the table(s) will be larger than the size of table shown in the pg_table_size result.

Therefore the size transferred is correct, but the size of the table is smaller as its compressed.