4 Samsung 850 Pro 1TB SSD with Qnap 879-Pro - SMART Error and Unrecoverable errors

bladeslap22

Reputable
Jan 7, 2016
7
0
4,510
I have a QNAP TS-879 for which I installed 4 Samsung 850 Pro 1TB SSD drives about 90 days ago. I am using VMWare 5.1 and had two rather small low utilization virtual machines that have been working flawlessly until this morning. The VMWare server reboot (which never happens) at about 5 something in the morning. When I went to start the VMs, I received a SMART error from the NAS (Media Error)

Drive4 Read I/O error, UNRECOVERED READ ERROR sense_key=0x3, asc=0x11, ascq=0x4, CDB=28 00 03 f8 c2 88 00 00 80 00 .

Then 3 hours later I received the same exact error for Drive3...

I viewed the SMART status of each and it showed a problem. The relocated sectors at one and gave me a warning for both drives. I then got another media failure and unrecovered read error.

I contacted both QNAP and Samsung. First, I transferred all the data off the SSDs to my conventional drives and then took out the SSDs and ran the samsung magic utility (forgot the name) to query the SMART status. Attached is what I got for the two drives causing the problems:

ssdproblem.png


So, one reallocated sector count on each drive but they both failed the ECC Error Rate and Uncorrectable error count. The other two drives were fine. These were #3 and #4 in the RAID5 array.

QNAP said that they're 100% sure it's a drive problem...but it's odd that they happened within 3 hours of one another ... However I do have more disks in the QNAP that are not having any problems at all.

So, the array was blazing fast ... loved it, but am very concerned to try again, especially with SSDs.

Anyone care to share their opinion or shed any light? Thanks
 

rmhall

Reputable
Feb 3, 2016
7
0
4,510
FWIW: I just had a Samsung 850 PRO 2TB start exhibiting the same things with almost EXACTLY the same number of hours on it - 3543 (its up to 3573 now after troubleshooting, etc. the past few days). I suddenly had read errors, and it literally locked up- my entire system (used as a boot drive on a Mac Pro connected via an AttoTech ExpressSAS H608 adapter internally. I checked all my cabling, etc. and it was definitely the drive. The ECC Errors shot up to 369, same thing with the Uncorrectable Error Count. I also had 4 Runtime errors logged out of a threshold of 10 and 4 reallocated sectors. The ECC rate and Uncorrectable Error count are way past the threshold to indicate a FAIL according to Samsungs own Magician Software - but it still rated the drive as "HEALTHY" - ridiculous to have it start failing like this with only 147 days of usage and only 37 power cycles. I tried to keep using it, but it would still dropout and hang my system anywhere from a few minutes to a few hours of use once all this started happening. I contacted Samsung for support, and got an RMA, as I simply can't depend on the drive at this point and to have it already be in the FAIL and near failing for several other SMART status items is unacceptable on a drive that is supposed to have a 10 year warranty, and 300TB worth of writes - I only had a little over 8TB of writes in that time.

I was able to do the secure erase before I sent it in today.

It is really odd that it started failing right around the same # of hours as yours with similar types of failures - I am wondering if there is a pattern here - I am guessing you bought or at least started using yours back in August or so based on the power on # of hours?

I am hoping that my RMA is processed and I am sent a new one and don't have to argue with anyone about its reliability.

Good luck!


 

bladeslap22

Reputable
Jan 7, 2016
7
0
4,510
Okay there must be an issue in the manufacturing of the drives. I find it VERY hard to believe that my two drives and yours would fail at the same exact time. This needs to go to Samsung so they can evaluate what the problem is. I immediately pulled all the virtual machines off the array when it started happening because errors kept popping up. The entire VMWare server rebooted...it wasn't pretty.

The other odd thing is the hour count was not accurate ... I only owned them for 90 days before the problem started yet the number of hours indicated a month or more past that. Odd... Did your hours equate to the actual on time or did it seem a little longer?

I sent them back and got refurbished replacements. I sold the two good ones and then sold the refurbished ones as well - I was also going to use them in an enterprise situation and decided against it. I purchased the 845DCs instead for my client as they appear to be in a different league. If you froogle it, you'll see the prices are starting to come down quite a bit. Just my two cents
 

rmhall

Reputable
Feb 3, 2016
7
0
4,510
Sorry - got sidetracked dealing with Samsung - HORRIBLE SUPPORT.

I have to agree, its far to big a coincidence for the drives to fail in a similar fashion at nearly the same amount of power on hours. I found your post after I had already sent things in, but I am going to reference it in some follow up with them. You don't happen to remember or have your ticket #'s from Samsungs support that I could reference do you?

As for the # of hours being off - I think mine is fairly accurate - I received mine September 6th, 2015 and I believe I installed it that day, so it is pretty close, give or take a couple days - not sure what the value was when I first got it - going to check whatever I receive back from Samsung.

Thanks for the tip on the 845DC - if they had a 2TB version I might consider it - but it looks like the 845DC is only 1TB and already double the price of an 850PRO and I'm not sold it would be that much more reliable. I thought something with a 10year warranty and 300TB and a couple million hours of MTBF would last me - apparently I was wrong, disappointingly so.

So here has been my experience since I last posted:

Contacted Samsung via their 1-800-SAMSUNG support on Monday Feb 1st shortly after 9am through their phone system. I explain the problem to the service rep and asked him if they offered any form of cross shipment's or any ways to expedite repairs. His answer: no but it would only take 3-5 business days to process and repair.

I say OK, and get an email with the instructions to send in all the info (proof of purchase/invoice, photo of drive with full serial #, etc) - I sent that in a few hours later on the 1st. They make me wait over 24 hours till the evening of the 2nd (after 5pm) just to send me the UPS label and RMA # - I am unable to make the cutoff time for the UPS Pickup on the 2nd, so it sits in a UPS box overnight to get picked up on the 3rd for overnight delivery.

It arrives at the destination at 9:38am on the 4th (the support location is in Carlstadt, NJ just north of Newark and NY, NY on the way to Hackensack - I could have driven there in under 2 hours) and I am only in Philly, so it gets there quick and is signed for by "LEE". However, it just sits there - all day Thursday, all day Friday - through this past weekend, and all day Monday the 9th - it is not processed at all. The self service page that you can check your ticket/reference #, continues to display that it has not been received, hasn't been updated, etc.

Last night I get an email from Samsung, saying they haven't yet received my drive and to reference the ticket # they gave me, and TO CALL THEM WITH PROOF OF ITS DELIVERY. This is the same support that issued me the UPS label with the tracking information on it. Which they could have easily looked up and they would have seen it was delivered last week. I am beyond pissed at this point.

I start calling just before 9am this morning Tuesday the 9th, and after their system hung up on my at least 10 times until precisely after 9am - it takes about 10 minutes and I get through to the SAME guy who took my original call. I explain to him the email, and that it was in fact delivered to their location last week, and I ask why it hasn't been updated yet. He then informs me that "Yes you are right it hasn't been processed yet." I am getting madder by the minute. I ask when it will be processed since its been sitting there since last week, and I was told 3-5 business days from start to finish. He has no answer, and just says it will likely be updated soon. No explanation on why their automated system couldn't have looked up the tracking # with UPS that they assigned when they created the ticket and RMA last week.

Finally at 11am today I get an update from Samsung, that it "has been received" - but now the kicker is - it won't be processed until the 15TH! And that falls on a holiday - Presidents Day - so I don't expect to see anything until the 16th. TWO FULL WEEKS from the time I contacted them.

Samsung's support is downright horrible - they have outsourced it all to totalts.com and I have no idea what their relationship is - but its piss poor service. I can only imagine it was all outsourced because they are dealing with tons of defects or problems.

I am so disappointed at this point. If I get a refurbished drive back and not a brand new one, I am going to be pissed. At this point, I just want to sell it and never deal with Samsung again. Sigh... more than anything, its wasted my time more valuable than anything else. I need to find a good contact at Samsung to relay my experience and their horrible support system and staff.

 

bladeslap22

Reputable
Jan 7, 2016
7
0
4,510
I am literally lauging my a** off bevause my experience was IDENTICAL TO YOURS. Even the part where it took several
days to get the RMA to the point where they told me they had not reveived the drive when the UPS tracking number clearly showed otherwise. I just sold the returned drives on ebay and got the 845s - Check froogle as they are starting to come down dramatically. The reason the 2TB isnt available is the controller they are using is previous generation.

Also, from what I read, the 845s are in a completely different class. Will see because I have them installed in an enterprise application now.

Let me know how it works out!



 

Palorim12

Distinguished
Well the 850 Pros are not recommended to be used in RAID. So its at your own risk.

I've never had a sour dealing with him, maybe its all in the attitude you present yourself with them. My last repair which was around the end of January, i sent it off on the 29th, and got the repaired unit back on the 4th of Feb.
 

bladeslap22

Reputable
Jan 7, 2016
7
0
4,510
Not sure how our collective attitudes affect the initial RMA taking two days to receive ... And I had to call them up for them to manually send it ... then the repair facility did not acknowledge receipt of it until a few days later. All in all, it was a very tedious process with far too many steps involved. They need to streamline the way it works.
 

Palorim12

Distinguished


I meant attitude in regards to how the agent you spoke with talked to you.

Repair wise, i contact Samsung to start repairs on behalf of my customers all the time, and normally the repairs or exchanges are pretty swift. Maybe there's something going on in the last week or 2 that is causing them delays.
 

bladeslap22

Reputable
Jan 7, 2016
7
0
4,510
The person I spoke with was very cordial as was my demeanor with him. There is something wrong with the process in how they handle RMAs. I've been doing this a long time and never quite bumped into a company that makes you wait 2-3 days to receive a second email, for which you have to fill out, and then send in to then wait a few more days for an RMA. It was a disorganized mess, putting it lightly. That being said, it didn't take too long to get a refurbished drive back, so all in all, it was painful at the onset, but worked out at the end.

Also feel uncomfortable that I had two drives go (out of 4) at the same time and the other gentleman on here had is drives also go within an hour or two of when mine went bad (power on hours). That's disturbing to and may point to a bigger issue here.
 

Palorim12

Distinguished


Oh ok, i misunderstood the way you wrote that initially then.

I've always received the ticket confirmation and UPS label within 24 hours.

Well, like i said, the way you're using it is not recommended by Samsung, since they don't recommend or support RAID setups, so that could be part of the issue.
 

bladeslap22

Reputable
Jan 7, 2016
7
0
4,510
In my mind, data is data is data. The drive doesn't know if it's in a RAID configuration or not...Whether it is striped or used as the sole drive, it's really irrelevant. Samsung said they don't support it "because they have not tested it".

Now, that being said, the 845DCs are built to be much more bullet proof and they are certified/tested with certain controllers. So, when you are providing a solution for a client and it stops working, you have manufacture tested compatibility to fall back on ... I don't think having put them in a RAID configuraiton did anything to the drive ... many people have used the same exact ones with QNAP successfully. I think there was a bad batch out there and I think Samsung is aware of it.
 

Palorim12

Distinguished


If you look out there, there's tons of places that talk about why its not recommended to use RAID setups with these SSDs.

The 845 DCs are discontinued. They were replaced by the PM863 (successor to the 845 DC EVO) and SM863 (Successor to the 845 DC Pro)

http://www.samsung.com/global/business/semiconductor/minisite/SSD/global/html/datacenter/overview.html
 

bladeslap22

Reputable
Jan 7, 2016
7
0
4,510
Thanks for that info ... was not aware they came out with a new line. In fact, the price point for the 863 appears far better than the 845 series.

If you happen to find a link that adequately describes the downfalls of RAID on those drives, would be interested in researching it a bit more. Thanks again for your input into this.
 

rmhall

Reputable
Feb 3, 2016
7
0
4,510
Quick follow up - Yes, my initial demeanor was totally cordial as well - it was only after continuous delays in an overly onerous process (some of which I outlined previously and bladeslapp22 also experienced) and breakdowns in this self imposed process that have frustrated me. The latest is that their self service status page is just plain broken, and their automated text message system literally sends me back an error saying its broken - I mean how much more can break?

It's good to hear Samsung has some other more enterprise level products - but if its the same support you have to go through in the case of failures with those offerings - forget it - I'll go elsewhere gladly and will be happy to pay a premium for expedient and superior support. Samsung support process is totally unacceptable. Do they have a special enterprise level support that won't get routed to totalts? That would be a determining factor for me to give any more money to samsung on their products.

Also, my drive was not being used in a RAID, it was just a boot drive in a system that gets low usage, and was rarely power cycled, and yet failed in the exact same manner, and at the nearly identical number of hours as bladeslap22's - I also find that to be too much of a coincidence.

If anyone cares, I just got off the phone again with them (after having made initial contact on the 1st) and was told that now it will be well into next week if I am lucky, because indeed it was a failed unit, and cannot be repaired, but has to be exchanged with a refurbished unit.

According to the rep, this now allows them to reset the "time" of 3-5 business days and gives them more time to get back to me, as I had asked why the process was taking so long beyond the initial 3-5 business days I was told when I spoke with them originally. They took their sweet time in processing the unit once it was actually returned to them. Once they receive it, to me that sets the clock ticking, not when they get around to actually start processing it after receiving it. And who on earth had an expectation that they were actually going to "repair" an SSD drive - are they going to take it to a desoldering station and replace surface mount components and IC's on each and every one and re-certify it? No, they do some basic tests for confirmation of problems and send out a refurb or a new unit. Ridiculous answers and processes. I feel bad for the poor guy on the phone. I wasn't rude, but I was pretty curt with him today as I plead my case because this process is so painful, had it been for a client I would have been even more irate. I even apologized and acknowledged that I don't find him at fault, but that they should acknowledge the process has taken inordinately more time than I was told originally, and the continuous breakdowns have simply made things worse and more tedious.

If I had to hazard a guess, if there is in fact a systemic issue with the drives failing (maybe a bad batch) and even a small percentage of those are failing at the same # of hours, then they may be backed up from getting a good chunk of those all at once just recenty - that make make sense, but who knows, and no one from Samsung or totalts is giving me any information, just continuous delays and excuses. If they at least provided a bit more transparency and attempted to make a customer a repeat customer, instead of being argumentative, I might have a different taste in my mouth.

More than anything, the original failure cost me several days of hassle in swapping drives around, troubleshooting, restoring backups, properly securely erasing the drive while it continued to fail, that's fine - I expect failures, its not a matter of if, its when, so I am prepared with backups - but all the back and forth and time wasted trying to get the drive replaced, etc. - time is something I can't get back, but could be compensated for - Samsung hasn't stepped up and done that. Not good customer service - tons of room for improvement.

I plan on updating my review on Amazon where I purchased it, so that others will understand what they are up against if they do end up needing support. I can't in good conscience recommend a product like this where the support (if you are unfortunate enough to actually need it) is so problematic. End rant - the horse has been beaten enough...
 

rmhall

Reputable
Feb 3, 2016
7
0
4,510
FYI: To be a little productive, here is a review of the Samsung SM863 which is the enterprise level SSD drive optimized for write operations: http://www.storagereview.com/samsung_sm863_ssd_review

And the PM863 drive optimized for read operations comes in even higher capacities - review here:
http://www.storagereview.com/samsung_pm863_ssd_review

Thanks for the info on those drives Palorim12.

The SM863 does come in a capacity of up to 1.92GB for that model, but at an over 200% premium in price compared to the consumer level 850 Pro - I see it selling for around $1800-$1900 right now. Ouch! (Corrected, found much better prices closer to $1132 and $1032 for the PM863 - much better!)

The PM863 designed has an additional higher capacities, but the price point of the 1.92TB version is not too bad after doing some searching - its not that much of a premium over the 850pro - so interesting overall.

I wish the reviews also mentioned what kind of support you get and if it is a different tier of support with their enterprise products - will have to dig around for that a bit more, there is nothing on Samsungs site that indicates anything for service beyond what you would get for their regular consumer/pro-sumer level drives.


 

Palorim12

Distinguished
Well all the PRO and EVO drives are consumer level products, so Total tech handles all their Consumer level SSDs. The other stuff is Enterprise/OEM products, and all i know is that you have to go through a branch of Samsung called Samsung Semiconductor, but they only speak to Businesses.

I think you might not understand their turnaround time. I deal with them almost daily and have good repertoire with the agents there and this is how it was explained to me. Turn around time for Repair is TYPICALLY (Which is normally, not promised) 3-5 Business Days, Not counting weekends or holidays. If the repair fails, then they have to replace the unit, which extends from TYPICALLY 3-5 business to 7-10 Business Days. So keep in mind that is still from the moment you shipped the drive and once again, nomrally how long it takes and doesn't take into account delays or weekends or holidays.

SO, based on your information you provided earlier, you shipped out the drive on the 2nd...BUT, UPS didn't Pick up until the 3rd. So, due to dealing with them for the last 3-4 years, i know for a fact they use next day air shipping, so lets say they received it on the 4th. That's one business day so far. you called them on the 9th and they scanned it in that day in their system, so by that point its the 3rd business day. So today would be the 5th business day, so if you haven't received it yet, means most likely that the repair failed and they are going to exchange the unit. Which means there is still 2-5 business days left in the turnaround slot they tell everyone.

Since everything is actually going routinely, i think you're just being impatient. Also, seeing as you said you got your drive on September 6th, if you are going to receive a replacement, its gonna be a re-certified unit, because they only give new units for ppl that are within 90 days of purchase or manufacture date(only if you didn't provide some kind of invoice or receipt).
 

rmhall

Reputable
Feb 3, 2016
7
0
4,510
Good to know they have a separate division for handling their enterprise products commanding a higher price point.

So I have my own business, fully incorporated C-corp since the early 90's so no problems there - or are you saying they only speak to "Businesses" with a certain volume or amount of business, or OEM relationship? That is quite a different story and understandable if it's volume or some other requirement beyond being a "Business".

However, it looks like anyone, regardless of being a "Business" or not can purchase a PM863 or SM863 - are you saying if it's an individual and not a "Business" that you won't get the support through Samsung Semiconductor, they won't speak to you and you have to go through totalts instead? That doesn't sound right to me. Is there some other criteria for qualifying for enterprise level support beyond just the purchase of an enterprise level product?

I don't meant to be argumentative, as you have provide some good info on the other drives (I appreciate that) and it definitely sounds like you have had a different experience. I am sorry to hear you have to deal with this daily! I hope that doesn't also speak to the reliability of the drives in question and Samsungs SSD en masse, and just the volume in whatever you are involved in - yikes! However, for the record: I do understand turnaround times 100% and more importantly, I understand good customer service. What you are not taking into consideration is the extra time they interjected into the return process, that both bladeslap22 and I experienced. They made us both wait over 24 hours to send me the RMA information once I provided them all the proof of purchase, etc. Thats 1 business day I am still down. Then, when it did arrive (via UPS overnight with the tracking # they provided) it sat for over 2 business days who knows where in their facility, which is odd as the label included the department, RMA ticket, etc. Why do I have to call and inform them "Hey remember that RMA you issued me from your system, and the overnight UPS label you gave me, that I used to ship it with, it arrived 2 business days ago, you mind processing it now?" There was never anything indicated to me in the documents or on the phone with their reps that I would need to follow up and make sure they process the RMA once they receive it, and not when they "get around to it" - thats ridiculous. I've never had to do that for any product with any vendor I have worked with. In 100% of all the other support services I work with, I always offer to put up a credit card and allow them to put a purchase on it for the full amount and cross ship a replacement ASAP, and when they confirm it is indeed a failure, release the hold and we both part ways satisfied. Unfortunately that wasn't and option and there was no way to expedite their service. I also wan't informed that if it indeed was a failed drive or that they couldn't make an actual repair, that it would involve additional time - I was simply told from the onset it would take 3-5 business days from the time it was received till everything was processed. So lack of communication of details like that (again poor customer service) led me to make certain plans, that had I been aware of the details, would have led me to do things differently on my end - so again my time wasted.

So no, I'm not being impatient, I'm frustrated because: someone else wasted my time by not processing things like they said they would, that I had to babysit this from the start to finish, that their self service status systems were 100% broken for me and my transaction, and that I wasn't told in detail there would be time above and beyond the 3-5 business days if they couldn't do a repair and had to issue a replacement. Time is in fact more valuable to me than money, and I don't like it wasted. I know what good customer service is, and wasting my time is the polar opposite of good customer service.
 

Palorim12

Distinguished
When i say i talk to them almost daily, its not like i'm setting up multiple repairs with them everytime. Sometimes i call them asking about new products, or if i'm stuck on something and having a brain fart. If i were to average asking me for a repair, it would probably be like 1 or 2 every 2-3 weeks. And the place i work for is in a high traffic area, so we get alot of walk-ins.

I just know of the PM863 and SM863, because i read reviews about them when they were announced. idk, about who handles what.

Like i said, i've never had a bad experience with them, and due the frequency of me calling them, i know they had 4 or 5 guys in the dept, but lately i've been hearing some new names. I called got a guy named Isa or Isaac, or something something staring with "Is" last week who i'd never spoke to before there, i had to hang up and call back because most of his info was incorrect or slightly off. Called back 5 min later and got someone i've spoken to many times over the last couple years on that next call, and he gave me all the info i needed and apologized for the other guy because he's new.

We had a new guy not too long ago that just like that, that i had to train, no matter how many times i corrected him and told him the right information to tell ppl, he would somehow distort what i told him and combine his bad information with my correct information and make things worse. I have to apologize for him all the time. the only reason i can think of we haven't let him go is because we are shortstaffed and need everyone. if i could, i'd kick him out the door, cuz he causes more problems then actually helps.
 

rmhall

Reputable
Feb 3, 2016
7
0
4,510
You just nailed it - same guy I've been working with - I won't put his full name out here, as its clearly not entirely his fault (especially if he is new, we've all been the new guy at some point - so I'll gladly cut him some slack, especially if they figure out way to make me happy) - but the process is flawed (IMO) far beyond the poor souls who man the phones and capture the info. Anyway - no need to keep hashing this over - I had a bad experience, waiting for it to be done so I can get on with my life.

Got it - so if I can in fact get enterprise level support from Samsung where the service is commensurate with the premium in price (cross shipments, accurate info, quicker turnarounds, no delay in RMA generation, don't have to deal with totalts, etc) and the warranty is similar or better (as there is no mention of the warranty info on Samsungs site for the SM863 or PM863 that I can find - its still showing the info for 845's which although are not rate for 10 years, the TBW are much higher which is more important to me anyway) then I'd consider the SM863, as its not a huge premium over the 850 PRO in terms of price, and it is being represented as more durable.

Will see how things pan out next week when I get a replacement...
 

Palorim12

Distinguished
They used to do cross shipment. The repair first thing started back in the middle of January. I had called up and asked them why they changed. Thankfully i got someone who i know was there a long time, he informed me that its because "they were having issues where they would send the Advanced unit, they wouldn't receive the defective unit, so they would charge the credit card, and then the customer would just call their CC company and do a charge back, which basically would undo the charge, and also combine that with the fact that about 75-85% of the units they received in the exchange would be tested and come up with no defect and pass all their tests, they decided to switch to a Repair first policy."
 

rmhall

Reputable
Feb 3, 2016
7
0
4,510
That is unfortunate that they no longer offer cross shipment. It's a shame that they have to process so many repairs that the percentage of chargebacks is so exorbitant that they are unable to carry that as an expense of doing business in order to provide quality customer service equivalent to most other vendors. This also gives me pause that they haven't been able to work out common supply chain and support/credit card chargeback challenges that are by no means unique to Samsung. I am always confident when I need to send something in for repair, that it actually has failed and needs a repair or replacement, so I would have no issues with a restocking fee, or some other charges they have to process and test a device that has no faults, and passes. I know not everyone may be that confident, but that seems like a reasonable (and common) solution to the issue of processing drives that don't actually need repairs. In any case, thanks for passing along the info though - I appreciate it!
 

TRENDING THREADS