Drain response body on exec call to enable keep-alive connection reuse

[Fiery-Fenix]

Currently library generates big amount or errors on ClickHouse side:
ServerErrorHandler: Poco::Exception. Code: 1000, e.code() = 104, e.displayText() = Connection reset by peer
That's happens because of improper close of response body.
Documentation in net/http/response code states that to reuse keep-alive clients require to both read to competition and close body:
https://go.googlesource.com/go/+/go1.15.7/src/net/http/response.go#63
Unfortunately library only close response body without reading response body which leads to dropping keep-alive connections instead of reuse.
Also good discussion about this issue could be found here: google/go-github#317
Tested on ClickHouse versions 19.16.12.49 and 20.8.12.2

[DoubleDi]

Hi! Awesome work!
Is it possible to close the Ping Body too?
This fixes #72

go-clickhouse/conn_go18.go

Line 25 in e482166

respBody, err := c.doRequest(ctx, req)

[Fiery-Fenix]

Hi! Awesome work!
Is it possible to close the Ping Body too?
This fixes #72

go-clickhouse/conn_go18.go

Line 25 in e482166

respBody, err := c.doRequest(ctx, req)

Nice catch! Added