Need a pro to check these SMART readings.

S

SomeGuy133

2[H]4U
Joined
Apr 12, 2015
Messages
3,447
I know a large number of you know how to read the raw data on these SMART values and I wanted to ask if anyone could be nice enough to tell me if these drives are good. The 1 TB is super loud comparatively when being accessed but i checked the AAM or whatever and the acoustic one was set at performance (loud). So maybe the drive isn't bad but that's why i am hear asking. It is an older drive. The oldest I got actually besides the 10-500 GB drives i just through on 2 250 pound magnets a couple months ago (i don't have a deguaser yet...on Christmas list ^^).

Second side questions
Also I wanted to ask if there is a way to force a drive to be on performance settings permanently. I remember with the 1 TB drive years and years ago you had to reset the stupid AAM/APM(?) setting ever restart because it would go back to default. I remember it making a fairly large difference in regards to inner platter access times. They literally dropped to half the default latency if not more.

3 Drives SMARTs
https://www.dropbox.com/s/2cgrrwztk...d of movement-its loud AAM is ramped.png?dl=0
https://www.dropbox.com/s/o5jjijrrjne24zc/3 TB -1-.png?dl=0
https://www.dropbox.com/s/mratmujrf6pivg0/3 TB -2-.png?dl=0
 
I suggest you use a program that translates the raw hex smart values, such as smartmontools. Smart data is encoded in hex and the only data that could be "read" would be the parameters which report a decimal value, such as reallocated sectors. Thus you can do a simple hex>dec conversion. For other values such as temperature, the encoding will depend on the manufacturer of the drive which is where a program like smartmontools will know the correct way to decode the hex value.

As far as being able to give you a simple answer of "the drive is good/bad" isn't really possible when reading smart data. The typical indications people check for are anything to do with reallocated sectors and CRC errors. However, you could easily have a smart value of zero for reallocated sectors yet have some sectors that are bad but haven't been read/written in a long time (easily possible on large drives).
 
To add to what Camaro said, SMART implementations differ among brands.

The one with 20 000+ hours is clicking because it's pretty friggin old. Acoustic management does not cause periodical 'clicking'. But it definitely quiets down seek noise.

It's also important to view SMART data in context. The best way to utilize SMART data is to poll it periodically, with something like HDD Guardian, and watch for sudden spikes of parameters. That way you know what change is 'normal' and when it suddenly skyrockets you know something's up.

Also, again Camaro suggested smartmontools - a Linux thing. It's useful because it can read not only SMART data but dump the internal log. The drive itself performs self-tests and can note of a serious error and which commands lead to it. Typically it's something like 'READ DMA'. You can also force such self-tests (some are captive in the sense they 'block' access to the drive) and some are offline and you have to wait a few minutes till the results show up in the Smart log. MHDD is a great tool to trigger SMART self-tests.

An important thing to do also is to periodically launch something like bart's hdd test so the drive has an opportunity to scan normally unused surface.

While booting a Linux kernel, you can watch for a message 'HPA detected' or something like that. HPA is Host Protected Area, and normally it means the capacity of the drive has been reduced - because some of it is unusable.

So, as you can see, it's not as simple as posting one's crystaldiskinfo.

However. Based on my experience I'd say this about your three drives:

- 1TB Hitachi - looks quite old, clicking is never a good sign, and ULTRA DMA CRC is non-zero, meaning you need to double-check your cables, re-seat them. It CANNOT go up. A UDMA CRC error is a possible false read/write error. The Smart data is all over the place, but it's old, so...
- 3TB Seagate - quite new and yet the seek error rate parameter is quite high. I would watch the trend and how it changes over time. Also, again, 1 UDMA CRC event. This should be zero unless it happened like once while you were messing inside the case and it hasn't changed since
- the other 3TB Seagate - again, new but seek error rate is quite odd.

I can see that only the first 3TB drive was connecte directly while the other two via an USB adapter?
Reliable testing can only be performed by a direct connect IMHO. The adapter can screw up readings for a perfectly good drive. It's a point of failure. It can "lie".

All in all, Get something that plots graphs (like hdd guardian or speedfan IIRC) and start stress-testing by something simple like bart's stuff test. Then compare the graphs with normal usage.

They are not "perfect", hence my suggestion to test. Smartmontools will show you the smart log which is as important as the readings themselves.

Also, watch for disk related errors in the event log and use something like hashcheck to build a database of checksums for your files. Silent corruption can be thus easily detected.

Oh and as to forcing AAM into performance - again - a Linux distribution with a tool called hdparm will help you with that.
 
Last edited:
MHDD is also an important tool because it shows 'warnings' when the read time was oddly high - suggesting a sector that might be failing.
So, to fully test one drive, you need a few days of stressing and plotting graphs :D
 
the seek error might be cause i was putting these in an external minus one of them so it very well copuld be a cable issue. That is on my to do list of first checks.

Also all this linux stuff is sadly a no go for me. I only use windows. The linux stuff just adds more issues. :/

I'll have to remember to post keep this windows relevant next time :D

Although its nice to hear about these cool linux tools though. I'll keep it in mind.
 
It's quite easy. Honestly. I'm not some evangelist, it's just powerful as f**k.

All you need is a Live CD of a linux distribution, something like Slax - it has a graphical user interface and all that.

No risk of damaging your data.

Once you're up on a Live CD environment, and manage to find a terminal/console window, just let us know here. I can provide you with the proper commands and help you interpret results.

MHDD is not linux based, it can be found on Hiren's or similar boot CDs. It's ... amazing.
 
Seek error rate shouldn't appear as a typical USB anomaly. Ultra ATA CRC yes, some other stuff yes, but not that one.

If not Linux, lose the USB enclosures and start monitoring your SMART data with a graphing/plotting program like HDD Guardian (Windows).
Bart's stuff test is a cool little tool that simply writes to empty space and reads it back.
 
I'll just throw out that you can most certainly use smartmontools in a windows enviroment. I use it in both all the time!
 
Wow, even better Camaro

And you can smartctl -s on <device> -a ?

If so, how do you refer to the device?
Does it dump the log entries?
 
CamaroZ28 said:
I'll just throw out that you can most certainly use smartmontools in a windows enviroment. I use it in both all the time!
Click to expand...

yea but its this pesky command line stuff thats so much work -_- I am seriously lazy about learning more and more. I know after i learn it once I'll be good and it won't take long but i don't want to spend the time reading and installing and learning. Its more effort than wanted at the moment with other projects. <----thats the real big reason but being lazy is another:D

Thanks for the replies. I'll look more into all of these over the next couple weeks. The HDD guardian looks promising.

Also i meant ultra ATA CRC not seek. I missed that value so i am not sure what it said it was so i'll check it out and google a bit more on it. I'll take a good look at them over the coming weekend. I want to get 2 more 3 TB drives so I can do a 4 drive RAID 0 for media and games.
 
Ultra ATA CRC almost always mean cable errors. Like a poorly attached sata cable, or too bended, too long, etc.

The seek error rate deals more with actual mechanical work and it's quite bad for a new drive.

The former usually gets non-zero while in a USB enclosure, but the latter usually does not.
 
I figure the data might not be reliable at all. It's a sudden dip/jump that could at least prompt to have a look at event viewer. Nice graphs and sweeping :D
 
Last edited:
You must log in or register to reply here.
Back
Top