Potential Samsung F4 issues.

[Kaygee]

There is an advisory for the following models also now (Guerssing that Japanese=JP).

 

F4EG HD204UI/JP1, HD204UI/JP2

http://www.samsung.com/global/business/hdd/faqView.do?b2b_bbs_msg_id=387&page=1&faqType=QI&faqKeyword

 

HD204UI - "512B"

HD204UI/Z4 - "4K Sector"

HD204UI/UZ4 - "4K Sector with 512B Emulation"

 

AFAIK the extension codes don't relate to different types of drive but location of destination market and warranty codes.

I think the UI is the base OEM model, 3yr warranty.

Z4 is 1 yr warranty and

UZ4 is 2 yr warranty.

 

It could be the other way round.

[jbrodriguez]

i have one f4 in my array, so i did the following

 

1)

hdparm -W 0 /dev/sda

 

2)

Stop the array

un-assign the F4 disk

Start the array  (this will cause unRAID to forget the serial number of the F4 disk so it can be used as its own replacement)

Stop the array

re-assign the F4 disk

Start the array.

Let it re-construct the F4 disk.  It will fix any of the "data blocks" that were never written to the disk.

 

 

it's about 14 hours into the data rebuild and it's working at approx 7,300 Kb/sec. guessing it has to with write caching being disabled.

 

that gives me an etf (estimated time of finish of Saturday.

 

i was wondering .. since i have a spare wd ears, shouldn't i pre-clear the ears, then data rebuild on this drive where write caching works fine and should give me much faster write speed.

 

would it be ok for me to stop the data rebuild then repeat step 2) above, just switching drives ?

 

thanks in advance

[jbrodriguez]

im 57% into the data rebuild ... should be done at around 8am (est) saturday.

if data rebuild is successful, i'd guess i'd still have to flash the new firmware into the f4 drive and re-enable the write caching flag, in order to achieve better write speeds.

[jaj08]

Blah wouldn't you know it, I have 1 of these drives in use, and another drive going through my 'break in' process.

 

Taking array offline and I think my plan will be to move all the data off manually, remove from array and allow Unraid to rebuild the array without the drive. I will then Update Unraid to 4.6.1, Update the Samsung Firmware, and use Preclear_disk.sh with the -A advanced format option before adding the drive back into the array.

 

Does the above procedure make since? I didn't bother with the advanced formatting before as it was my understanding it just caused a performance hit. But If I am starting from scratch with this drive I thought it made since to get all the above done.

 

 

[Joe L.]

use Preclear_disk.sh with the -a advanced format option before adding the drive back into the array.

The "-a" (lower case "a") option to the preclear_disk.sh requests the OLD sector 63 alignment.

The new "-A" (upper case) option to preclear_disk.sh requests the new sector 64 4k-aligned start for an unRAID partition.

 

Please fix your post, since a new user of unRAID might read it and get confused. 

If you want a 4k aligned partition, use the "-A" option (upper case "A")  Do not use "-a".

 

If you specify neither "-a" or "-A", you get the old sector 63 non-4k-aligned partition start.

 

Joe L.

[jaj08]

Fixed, I also was going by the Wiki that said 4.6.1 supported advanced formatting, I assumed that was a final build but looks like 4.7beta would be my only choice.  So I guess I may hold off on the advanced formatting anyway as don't really think I want to run a beta.

[BRiT]

The version was going to be 4.6.1 beta then 4.6.1 release candidates then 4.6.1 final, but LimeTech opted to change the version number to a full 4.7. So unRAID is currently in the 4.7 beta stages, to be followed up with a 4.7 release candidate and then a 4.7 final. Of course it might be possible to go straight from a 4.7 beta to a 4.7 final if no issues present themselves.

[jaj08]

Got my 2 drives updated today. Took a couple workstations to find one with a SATA controller that the Samsung utility liked.

 

Now to just hope this really fixes the problem. For grins I am going to remove all data off the one active Samsung drive I had, and will cross my fingers that 4.7 goes final soon so that I can have these format properly.

 

 

 

 

 

[umax]

So, any more info if this drives, with unraid v4.7 & updated FW, are now safe to use? I have installed one of this a few days back (updated FW and precleared with "-A". Preclear didn't show any errors or relocated sectors. The speeds are on par with WD EARS drives that I have.

 

What are the test I should preform in order to be sure everything is OK with my data?

[Orbi]

So, any more info if this drives, with unraid v4.7 & updated FW, are now safe to use? I have installed one of this a few days back (updated FW and precleared with "-A". Preclear didn't show any errors or relocated sectors. The speeds are on par with WD EARS drives that I have.

 

What are the test I should preform in order to be sure everything is OK with my data?

 

I've been using two HD204UI actively in my system for 2 weeks now. One as parity and the other as data drive. I'm still keeping my old parity drive around for two weeks before I redeploy it.

 

I extensively tested the two drives moving 500+ GB data around fand back rom my windows machine to unraid using teracopy. TC automatically calculates checksums. While moving data, I was performing smart test on the drives and no checksum returned incorrect. In my case I'd say that the firmware patch application was succesful. I'm pretty confident in the drives at the moment.

[jaj08]

I have 2 drives in use as well but the thing to keep in mind is that the firmware bug has a very specific cause. Your not going to know if your drive is taking errors until you do a parity check.

 

From what I understand is if you don't use any addons, specifically Unmenu, then the cause of the bug should never occur. Its not a fault in unmenu either, its the fault of the hard drive not liking to be accessed by certain utilities during data copies.

 

If you do use unmenu, like myself, then I plan to avoid ever accessing it when data is being copied to the system.  Even with the firmware upgrade, and my planning to not use Unmenu during data copies, I personally would still never trust one of these drives as my Parity drive.

[umax]

Thanks for info guys. I dont use unmenu (yet!) and its not a parity drive either. So I hope I should be OK. I will run parity check after I move whole WHS server to unRAID, which will take a week or two.

[pras1011]

The problems with this hdd has been fully resolved with the new firmware. So lets stop going round and round in circles!

[jaj08]

The problems with this hdd has been fully resolved with the new firmware. So lets stop going round and round in circles!

 

 

Normally I would fully believe that, but the one major thing I find frustrating is that the firmware version doesn't change when you apply the patch. So your left assuming that all went well with the upgrade with no way to verify.

 

With that said I still trust the drives, which is why I have 2 of them, but with the lack of verifying that the firmware has updated I personally would never want this drive as a parity drive.

[ratmice]

Newbie here. I am about to build a new system, and one of these F4 drives came bundled with one of my components. Sorry to dredge up the old thread. I am hoping that the drive has a date of Jan 2011 or later, but if not:

 

•Can you turn of write caching just for this disk, or do you have to do the whole array?

 

•Since all is new, If I do turn off the write caching, I shouldn't run into the problem, correct?

 

•Would it be better to use this as a data or parity drive? Seems like a forced parity check would rebuild the parity drive without corruption, whereas if the data drive were corrupted, depending on when the last parity check was done, you could also have bad parity data, too.

 

I may have trouble updating the firmware on the drive, no real access to any DOS system (I'm all Mac).

[Joe L.]

Newbie here. I am about to build a new system, and one of these F4 drives came bundled with one of my components. Sorry to dredge up the old thread. I am hoping that the drive has a date of Jan 2011 or later, but if not:

 

•Can you turn of write caching just for this disk, or do you have to do the whole array?

 

•Since all is new, If I do turn off the write caching, I shouldn't run into the problem, correct?

 

•Would it be better to use this as a data or parity drive? Seems like a forced parity check would rebuild the parity drive without corruption, whereas if the data drive were corrupted, depending on when the last parity check was done, you could also have bad parity data, too.

 

I may have trouble updating the firmware on the drive, no real access to any DOS system (I'm all Mac).

Do NOT trust the drive in your array until the firmware is upgraded.  All you need is a boot cd or flash drive that can boot to a dos environment. I doubt if you need windows to flash the firmware.

[johnodon]

I apologize if this is an ignorant question....but...

 

I bought one of these and was going to RMA it.  But, is this only an issue in Linux?  If so, I may just throw it into the WD MyBook case that is now empty and use it on my WIN7 system.  I'd rather do this than take a $12 restocking fee hit.

 

TIA,

 

John

[Joe L.]

I apologize if this is an ignorant question....but...

 

I bought one of these and was going to RMA it.  But, is this only an issue in Linux?  If so, I may just throw it into the WD MyBook case that is now empty and use it on my WIN7 system.  I'd rather do this than take a $12 restocking fee hit.

 

TIA,

 

John

No, it is an issue in any operating system.

[johnodon]

God I hate to send this drive back.  Is the general consensus at least that the new FW works?  I already flashed the drive.

 

John

[dyrewolfe]

God I hate the send this drive back.  Is the general consensus at least that the new FW works?  I already flashed the drive.

 

John

 

fyi. I sent 4 750GB F1s on 01/26 and after being told the RMA process should only take 2 or 3 days, I'm still waiting. I inquired a couple of days ago for an update and was told they didn't have any replacements in stock to send and couldn't give me an ETA on when they would be available. Personally, I'd take my chances with the drive you have. I own one and it hasn't given me any issues. 

[drf]

The problems with this hdd has been fully resolved with the new firmware. So lets stop going round and round in circles!

 

 

Normally I would fully believe that, but the one major thing I find frustrating is that the firmware version doesn't change when you apply the patch. So your left assuming that all went well with the upgrade with no way to verify.

 

With that said I still trust the drives, which is why I have 2 of them, but with the lack of verifying that the firmware has updated I personally would never want this drive as a parity drive.

 

That's how Samsung explains why they did not change the firmware number:

http://hardforum.com/showpost.php?p=1036842006&postcount=366

 

In order to verify that the patch works you have to recreate the conditions for the bug to appear. It is written how to do that in that thread:

http://hardforum.com/showthread.php?t=1566067

 

[Joe L.]

Sorry, but that is totally irresponsible on their part.

 

They are asking you to believe it would take too long to get approval for a version number change, but they were able to offer a un-approved change to the firmware, apparently ALSO without approval, and expect you to go by a supposed manufacture date to determine if your disk MIGHT have the fix, since you have no outward indication otherwise, even after the patch is applied.

 

I still say, STAY FAR AWAY from these drives until they have a new firmware version number as distributed or a new patch that does change the version number.  Going by a manufacture date is not enough.  Pressure the manufacturer for a proper patch.

 

Joe L.

[Rille_lkp]

The latest patch for these drives changes the firmware version but looks like Samsung have messed up the list for which drives it is for. See my posts in http://lime-technology.com/forum/index.php?topic=10521.0.

[drf]

The latest patch for these drives changes the firmware version but looks like Samsung have messed up the list for which drives it is for. See my posts in http://lime-technology.com/forum/index.php?topic=10521.0.

 

It is not for our drives (it is supposedly for the drives sent to Japanese market, hence /JP in the model number). I agree that Samsung is irresponsible regarding that issue (and other issues). What's more testing WD and Samsung disks surface with MHDD or HDDScan shows that Samsung drives surface is of poorer quality.

[Rille_lkp]

Ouch, you're right 

 

My drives are working fine and found one guy who did the same as i did and emailed Samsung about it. They said it was ok but i've emailed them as well to make sure.