If you have a non-TLER drive (like a green) and it hits a bad sector, the entire server will stall for 2-3 minutes while the drive tries to recover the sector.
It could be up to 2 minutes as it tries to recover. If it figures out quickly that it's in fact a bad sector, a read failure can be issued much faster.
TLER is only relevant in hardware RAID. The issue is way, way overrated. It's irrelevant to out of the box NASes (they use software raid), other software RAID and also intel RAID.
I wouldn't say it's irrelevant. It is non-catastrophic to not have settable SCT ERC on a drive in software RAID, because software RAID unlike many/most hardware RAIDs, don't have a time out that causes a slow to recover disk to be ejected from the array. But a 2 minute stall is still not great.
Also, the manufacturers make it rather clear that they don't want the least expensive drives used in anything but RAID 0 and 1. The WDC green, blue and even black spec sheets say this in small print.
So it's a feature you don't need and hence why pay more?
It's not the only feature the spec sheets tell us about. Until someone tests them more thoroughly, it may be marketing fluff. But the Red's are expressly for 24/7 use, consumer greens aren't. Red's have a longer warranty. They supposedly have lower vibration and higher tolerance to vibration. I wouldn't call it nearline, but it's an interesting produce, has better performance numbers and lower power consumption than Green. So if anything Green needs a refresh.
Plus, lower SCT ERC time, even on software RAID, means that the RAID fixing bad sectors happens more quickly before they are left to fester and be subject to the slower firmware recovery. The faster the drive gives up, issues a read error, the sooner RAID will rebuild the lost sector data from parity on-the-fly and then cause reconstructed data to be written back to the faulty sector forcing the firmware to either write it correctly, or if a persistent write error is occurring, to remove that sector from use in favor of a reserve sector.
A green drive on the other hand is pretty reluctant to issue read errors, which isn't necessarily a good thing for the long term health of an array. (But then so are bad sectors - hence why regular scrubs and smart tests should be run.)
I would not even bother with RAID. I assume this is for home use? RAID 1 is not back-up. It's for preventing down-time on disk-failure which is irrelevant for home use. Just do a regular back-up to external HDD.
It's a reasonable criticism. RAiD is about improving uptime, and reducing the need/likelihood you have to restore from a backup. It is definitely not a backup or a substitute for backup. There are any number of tools that can do scheduled syncs from primary storage to a secondary storage or backup.
I have 3 green drives and they all work perfectly fine. Use them for storing mainly hd content (movies, TV series).
The history of these drives is really variable. The same model has different firmware and different platter numbers and different arial densities. So they're in a sense different products despite the name.
I recommend a regular ATA secure erase to zero the disk, forcing bad sectors to be located and removed from use. And also regular 'smartctl -t long' tests, coinciding with analyzing the attributes of the drive with 'smartctl -a' to make sure the drive isn't having problems long before you start seeing them with lost or corrupt data.
Corrupt data is a bad experience, even with backups. You can get corrupt files successfully propagated throughout your backup without much difficulty.