Changes an incorrect .Read() to an io.Copy()

[pboothe]

Fixes #69

---

This change is 

[coveralls]

Pull Request Test Coverage Report for Build 355

• 10 of 10 (100.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 100.0%

---

Totals
Change from base Build 353:	0.0%
Covered Lines:	482
Relevant Lines:	482

---

💛 - Coveralls

[stephen-soltesz]

Reviewed 2 of 2 files at r1.
Reviewable status: complete! all files reviewed, all discussions resolved

---

tarfile/tarfile.go, line 173 at r1 (raw file):

	// the risk of unreadable files. If too-large files become a problem, delete
	// everything up to the declaration of the header and then replace the
	// `io.Copy(t.tarWriter, contents)` line with `io.Copy(tarWriter, file)`.

Anecdotally, when we were running with pcap collection in sandbox one of the patterns we saw was a surge in memory in packet-headers followed by a surge of activity from pusher. I got the impression that packet-headers created some big files for high-speed clients. Just to say this may be a recurring theme during the load testing.