More hard disk discussion...

So, following on from my previous discussion…

What are everyone’s thoughts on drive manufacturers and drive type - Eco drives for example.

I read a lot about products before I buy, and I’ve read positive and negative stuff about pretty much every drive and drive manufacturer going.

I’ve just read this thread on qnap.com - http://forum.qnap.com/viewtopic.php?f=182&t=34201

Obviously the native RAID constraints for drives don’t apply to a Drobo unit, but what about features of drives just like the Eco drives which try and park the heads to conserve power/energy…

I use WD20EARS drives in my Drobo after having been a staunch Seagate Barracuda customer. I had a few Seagate’s fail within a year and decided to look around - the WD’s seemed to be about as good or bad as any other, but came with both decent length warranty and advanced warranty replacement service. I’ve since had 2 failures in the last 12 months - so no different from the Seagate drives really…

So - I’ve mentioned the one feature - the head parking of the “Green” WD drives - I’ve read the issue here is that they are prone to issues after a few months and that drive units which like drives to never go in to power saving mode (i.e. stop spinning).

One thing I’m also keen on - do DRI test drives with their products? I’m assuming so, and so is this information available to the public?

As I look at purchasing a couple more drives I wonder if I should try yet another manufacturer, especially after colleagues recommend the Samsung drives. Keen for everyone’s opinions and for an open discussion that everyone can hopefully benefit from.

Snapper

i think the drobo’s arent affected by the parking heads issues - as long as the drobo read/writes regularly enough then the head wont be parked all the time.

im not a linux guy at all but from what i read the issue was that linux stores writes then bursts them out to the disk, so the head parked every 8 seconds (the WD default) and the heads had only just gotten parked when linux would issue another write command, so the heads were being loaded/unloaded about 6 times a minute (they are rated for 300,000 cycles)

drobo/pro/elite dont run linux so i dont think they are affected.

of course you can always either change the timeout for parking the heads or simply disable that one feature altogether if you are concerned (i’ve simply disabled it on all of mine)

i never really worried about it since the WD GP drives are the ones DRI chose to bundle with drobos so i will give them the benefit of the doubt and say that they have (hopefully!) tested them pretty thoroughly!

the saumsung drives are ok - i had the F1s in my desktops - they were very fast but a bit warm

seagate - after i lost data because of their faulty firmwares, im never buying another seagate drive (plus their currently 3TB drives are heat emitting monsters! - read anadtech’s review of the external unit - 70C…)

i did have 4 x wd gp and 4 x hitachi in my drobopro - the hitachis simply because they were VERy cheap when i got them - and they never caused my any issue

(aside from one DoA… i plugged it in and from the sound of it , it must have had a HUGE knock during shipping and it appears to be dragging its heads across the platters… made the most horrific screaming sound (i dont think i’ve ever unplugged a product so fast in my life!))

i’ve recently purchased 10 x wd20ears, one died within a week, but the rest are chugging along quite happily

I just had Drobo mark one of my WD20EADS bad, so I’m getting a replacement this evening.
If someone can point me to a utility to read the load cycle count, I can check the drive - this drive has been in a Drobo v2 its whole life and never had the Idle3 timer changed from its default.

SMART-wise it’s still OK, so I believe the counter should still be valid, even though Drobo doesn’t like it anymore.

its in the smart info bhiga -i sent you a screenshot of my SMART readout showing the one drive which was off the scale?

Yeah, will use yours as reference, I think DLGDiag can read the info to a file… if not I might have to download a util to pull the SMART data.

row 193 :)[hr]
errr

ive jut run DLG for windows- you only have two options - run test or view smart info - if i click view smart info i get a table with the load cycle count right there mmediate[hr]
in case you were wondering the load cycle count on my intel SSD is zero :stuck_out_tongue:

That’s what I get for using the DOS util, I guess, heh.

Grrr… SATA and I are having a disagreement.

I connect the drive to my Win 7 x64 desktop, no SMART.
Connect to my XP laptop, no SATA, and via USB-SATA, still no SMART.
Connect to my Win 7 x64 HTPC, no SMART.
Connect to my WHS machine, no SMART.

What the heck?

maybe that drive isnt smart?
(remember a tv show called “get smart”? :slight_smile:

it would be ironic if the only thing which has failed is the failure reporting mechanism

The SMART test passed in DLGDiag for DOS (both 5.04f and 5.19) - just the WinDLGDiag isn’t giving me the option to view the SMART data.

SMART Status reports Good in other apps, but I don’t know if the Good/Bad is just a flag independent of the individual statistics. It would be pretty ironic if the “dump the SMART table” function is what’s broken, LOL!

Second run of Full Test reported no errors (which tells me it remapped some bad sectors to spare sectors).

I think on my desktop (7 x64, AMD) it’s because the SATA controller might be in RAID mode.

On laptop (XP x86, Intel Mobile P4) issue is that it doesn’t have built-in SATA, and WinDLGDiag probably doesn’t like the USB-SATA converters I tried.

HTPC (7 x64, Ion), I don’t know why it can’t pull the SATA info.

WHS machine (WHS v1, Ion), also don’t know why it can’t pull the SATA info.

I brought the drive to work and will ask my friend here if he can hook it up and get me the info. I’m not giving up this easily! :slight_smile:

…and I do remember Get Smart, the shoe phone hehehe. :)[hr]
Okay, thanks to my friend here whose Dell lappy has ESATA, here’s what we got from WinDLGDiag and HDTune.

Load Cycle Count
Value 130 (DLG + HDTune)
Worst 130 (DLG + HDTune)
Status OK (DLG + HDTune)
Raw 210421 (only in HDTune)

It also has 148 reallocated sectors now.

Eeek, makes me wonder if maybe I should change the Idle3 timer??
The whole Value vs Raw is a mystery to me.

yeah its confusing :slight_smile:

ive got this in HD tune (for another internal drive)
Seagate Barracuda ST340014A 40 GB 7200 RPM IDE

SMART:

HD Tune: ST340014A Health

ID                               Current  Worst    ThresholdData       Status   
(01) Raw Read Error Rate         61       59       6        99128401   Ok       
(03) Spin Up Time                   98       98       0        0          Ok       
(04) Start/Stop Count              100      100      20       11         Ok       
(05) Reallocated Sector Count    100      100      36       0          Ok       
(07) Seek Error Rate                87       60       30       627347869  Ok       
(09) Power On Hours Count        52       52       0        42237      Ok       
(0A) Spin Retry Count              100      100      97       0          Ok       
(0C) Power Cycle Count           100      100      20       439        Ok       
(C2) Temperature                   32       41       0        32         Ok       
(C3) Hardware ECC Recovered      61       58       0        99128401   Ok       
(C5) Current Pending Sector      100      100      0        0          Ok       
(C6) Offline Uncorrectable         100      100      0        0          Ok       
(C7) Ultra DMA CRC Error Count   200      200      0        0          Ok       
(C8) Write Error Rate               100      253      0        0          Ok       
(CA) TA Counter Increased        100      253      0        0          Ok       

Power On Time : 42237
Health Status : Ok

interestingly DLG diag is a bit rubbish - i have two drives in my system - it can only see the one with a visible partition on it - the other one … the disk simply doenst show up ! i know its working fine because its full disk truecryted and its currently in explorer and working perfectly !

Yes quite weird. In my case the drive isn’t mounted by Windows since it came directly out of the Drobo, but so finicky this stuff is…

Just checked the SATA card on my desktop and it’s not in RAID mode so should have worked. shrug

SMART makes me feel dumb!

yeah it think the win DLG only works if it sees a lovely NTFS/FAT partition on it - it doesnt even see my truecrypt drive!

Well, it was happy with on my friend’s machine (where I got the SMART info) so I guess it’s not totally dependent on having a Win-happy partition… but it still is dang finicky.

No wonder the folks with Drobo S are having troubles. SATA, I love you, but you craaaaaaazy!

no DRI devices, afaict run linux to talk to the disks themselves. I believe there is a QNX layer that does the actual device control on all of them. Where Linux is present, it runs on a separate processor, either external (droboshare), or packaged together on a dual processor system (the other ones that can run Apps.)