Understanding the S.M.A.R.T. values from a Synology 1812+ device

Question

I have a Synology 1812+ NAS with 8 3TB drives configured as RAID 5. Its running DSM 4.1. It was purchased to replace USB drives, consolidate storage and short term OS X backups using Time Machine. The device and drives are only 2 months old.

Every other week I started to get IO errors from two of the drives. The logs has the following error:

Read error at internal disk [3] sector 2586312968.

And later on

Bad sector at md2 disk3 sector 250049936 has been corrected.

The sectors never match. The recommendation is to run a extended S.M.A.R.T. test on the drives. I did and this is the values I got:

I then ran an extended extended S.M.A.R.T. test on one of the drives for which I received no complaints and here is the values I got:

The values look very similar. It is unclear to me if there is a problem and if not, what is the point of a S.M.A.R.T. test if it doesn't reveal any real problem? How should I then interpret these results and when should I know its time to replace HDD?

Solution

the raw data column usually represents the number of events happened. E.g., number of read errors in the first row. However, your numbers are so high, that I assume you have a Seagate drive, which always reports abnormal high raw error values (also when hard drive is OK).

What else you can see - Status column. It is OK for all parameters, which means exactly the same - your drive is generally OK.

As written at http://www.linuxjournal.com/node/6983/print, the VALUE column presents a current "normalized value", which should be always greater than threshold.

So your SMART data shows that all drives are OK. However, if you get a lot of read errors (not just one found in the logs for last year:), it seems that your drive are going to die soon. It's somehow "normal" to have several (up to 1-2 thousand, see How many SMART sector reallocations indicate problems?) bad sectors on the drive which will be replaced with other and therefore corrected. But if you have too many of such messages or they come very often, you should replace your drive.

You can probably make SMART tests or some other tests (both depends on your NAS)... E.g., if you have smartctl and can login into NAS via ssh, you could try:

# smartctl -t short /dev/<device>

This command will run a short test for selected drive. After it will be finished, you could veiw results with

# smartctl -H /dev/<device>
# smartctl -l selftest /dev/<device>

Comments

[bloudraak]

I do have Seagate drives. The drives are relatively new and the messages have stopped, but it is usually a different drive. The Synology 1812+ device runs smart tests once a week automatically and notifies me if there were problems. The links were very helpful.