After some time, I can confirm that the raw read error rate value isn't very meaningful nor reliable. I have seen this counter reset to 0 more than once since I posted this question. And although the normalized value has decreased, it is still way above the threshold.
The reason why I posted this question was because I started to realize that the HDD was noticeable slower than before. Opening the file manager could take even seconds, as well as opening other programs or files. And I could hear the HDD "struggling" and making noise as trying so hard to read just a couple of bytes (and could see the raw read error rate increase a lot just after any of those operations...). Formatting didn't help with that anyway. Ensuring the HDD was correctly connected and the connector was clean didn't help either.
Why was this happening? I don't know, but I ended up buying a small SSD to replace it and use the HDD as a secondary drive (for data only). That increased my productivity at work and my health too... :-P (hearing my HDD struggling for reading data was making me sick)
As a conclusion: the raw read error rate is not meaningful by itself. The HDD seems to be working fine still, and it will probably continue to work for another couple of years, although it continues to be slow (something that does not bother me while the OS is on another disk).
My personal recommendation: always make backups even when your disks are healthy and brand new. If you feel your HDD is getting slow, consider buying a SSD to put your OS on it. They are pretty affordable nowadays and, if you use your computer a lot (i.e.: you work full time with it), then you'll really notice the difference.
It may be simply a loose connection. Try to disconnect the HDD and install it again. – gronostaj – 2015-01-21T22:15:32.777
@gronostaj: The HDD is mounted with screws, but anyway I will try that. Thanks! :-) – Peque – 2015-01-21T22:29:39.820
1
There has to be some type of data connection. 215 million errors is not normal. Be sure you understand how S.M.A.R.T should be read it snot always obvious.
– Ramhound – 2015-01-21T23:37:46.667The raw value isn't very meaningful by itself, and not very reliable to predict any failures. If the normalized value is above the threshold, you're fine. Keep your backups up to date, that's all. – Pedro Werneck – 2015-02-21T15:43:17.673
@PedroWerneck: I forgot about this question. I've updated it with my own answer to "close" this topic. :-) – Peque – 2015-02-21T19:45:47.800
1215 million errors is perfectly normal and an inherent part of how modern hard drives operate. Most drives will however hide all errors from you if they can correct them - Seagate is the notable exception here. The raw value before error correction is an internal performance metric used by the hard drive firmware and not meant to be of value to the end user. That said, if you have access to the detailed technical specifications for the drive, you can draw meaningful information from it but it needs to be refreshed every second because counter resets in a matter of minutes. – qasdfdsaq – 2015-10-18T20:06:48.927