When doing a psql copy to csv by selecting from table . we are seeing high write iops compared to read iops . What causes the high write iops when writing to causes file instead of db?
The most plausible explanation for heavy writes ocurring during seq scans is the hint bits feature.
From https://wiki.postgresql.org/wiki/Hint_Bits :
PostgreSQL's MVCC mechanism provides a lot of useful features, but there are some confusing side effects of the implementation. One revolves around hint bit processing, which can result in heavy writes to a database table even though you're just reading from it.