Raid 5 failed. 3TB node dead. Recovered partially and going through GE thanks to forum

Node Operators
81

Sounds reasonable enough. 3 failed drives over 2 years is probably annoying, but not totally unreasonable. However, my own experience with Seagate drives has not been like that.

I had 5 or 6 of the bastards fail on me over the course of a couple of weeks, and then at least 2 or 3 additional drives over the next year, followed by the semi-regular dying off of the rest of them. It obliterated my trust in the brand. I think I have 2 or 3 of them left today, and I’m only using them for data that I can easily afford to lose.

In the meantime I think I have lost a single drive from a different brand, and that one was so old it was measured in GB, not TB.

Of course you’re free to trust the brand if you want. I’m not going to tell you what drives to use, I’m only telling you what drives I’m not going to use. Be glad that I’m not buying Seagate drives, that only means there is less competition in the market for you when you’re buying them.

2 Likes
82

With those numbers, I’m going to guess they were ST3000DM001 drives. I’m not going to claim Seagate never made a bad model… they did. This model was so bad that something like 80% would die within 4 years. Clearly that was only a very specific issue with this model.

You’re missing the point. It was 100% of my WD drives that failed. I have a total of 16 HDD’s in use. I’m guessing you have quite a bit more.

The only thing I can advise is to never buy multiple from the same batch if you can avoid it and when possible buy different models. At least that way the chances of them dying all at the same time are much lower. This isn’t necessarily practical for everyone though, but it works for me. I tend to expand one by one anyway. But I realize the only reason I can do that is because SHR2 from Synology allows you to mix drive sizes, so every single disk I buy gets me extra space. I never have to replace the entire array at once.

83

I was, but then again I was also joking :wink:

84

…ahem, Synology :wink:

1 Like
85

Like it never happened… thanks :wink:

86

Yes, it was ST3000DM001. A bunch of them died over a few weeks at my job, nearly taking a 6 drive RAID with them. As soon as a drive went down, we got a replacement under warranty, and I rebuilt the RAID. The only problem was that the warranty replacement drives were just as short lived as the original drives.

I don’t remember exactly how many times I rebuilt that RAID, at some point it started to blur together. Then one day, the drives started dying faster than I could rebuild the RAID, and we lost a lot of data. Not super-critical data, but we did store it for a reason. So I said to myself, getting these drives replaced under warranty is doing us exactly zero good. It a massive waste of time, and a big risk, and an avoidable one at that.

At this point in time, Seagate still did not acknowledge that there was a problem with these drives. Actually, I’m not sure that they ever did.

So anyway, we ate the loss, took the Seagate drives out of operation, and bought WD drives to replace them. We were a pretty small business at the time, so it was a noticable financial loss on top of the data loss, but we did not exactly have a choice.

Maybe one of the WD drives failed almost immediately, I’m actually not sure. Because if it happened, it was not a big deal, it was a one time event, and we were more than prepared for that.

All of our WD drives are going strong to this day. Well, at least I think they are. I haven’t gotten any alerts, and in the last few years we have moved all the important data to the cloud anyway, so I haven’t felt the need to check their SMART status for many weeks now.

I’m actually not dealing with many more drives than you. The problem was that in a 6 drive setup, we had significantly more than 6 drives in total fail over a relatively short period of time. It was utterly ridiculous, and Seagate didn’t even acknowledge that there was anything out of the ordinary.

I don’t need that kind of BS in my life, I don’t want that kind of BS in my life, and I’m simply not having any of it. Screw that. Even if Seagate makes completely passable drives today, I just don’t care. I will happily pay 20% more for a drive just to get a different logo on it. And if it makes me sleep a little better at night (which it does), then all the better.

2 Likes
87

Actually no, fingers crossed they won’t all fail before they become obsolete like their competitor. So 5 is the number to beat, so far no other have failed with an exception of one WD that I dropped while running.

Not sure what you all mean by “anecdotes”.

And by the way, HGST/WD cost about the same as Seagates for me, at places where I can get them for reasonable prices, if anyone’s got a better place, do suggest for everyone’s benefit.

88

I can imagine avoiding them after something like that happens. As for why anecdotes are meaningless, your personal experience with a few drives is usually not representative of all drives out there. However with this specific model it’s a different story, there is more than enough evidence out there that there was an actual problem with this drive. So your anecdotal evidence aligns with broader numbers.

1 Like
89

yeah it’s offensive to the users of the drives, i know it would cost them more and slow them down but some proper testing before launching what can best be described as a prototype or contaminated series.
after this happened with seagate i think backblaze introduce their vetting of new drive series…
they will literally buy up small numbers in the 40-100 drive range… to ensure the model performs to their standarts… lol

funny tho… one should think most manufactures would have the common sense to test their product…
so yeah i might not have lot more than a few drives to seagate… but i still lost data and the bad drive was good for like a week from when it came out of the box…

90

I think there are way too many use cases and environments out there to properly test every single usage pattern etc

91

yeah but one couldn’t really blame that in this case… i owned the damn thing for a week and not like it was production usage or anything i copied a bit of data to it and it broke…
and apparently that was common… meaning seagate didn’t test one bit… not in the slightest or they would never have pushed such a poor product

92

I had 3 different Seagate models fail. Two 1TB drives, two 1.5TB drives and one 8TB drive. No other drives have ever failed for me. Luckily there was no data loss from any of these. According to Backblaze stats all models of Seagate are worse than all models of any other manufacturer. I’m not saying nobody should buy Seagates; for some it might be more economical, I’m just saying all are less reliable according to available stats. If the TCO for you is better with Seagates, go for it, that’s mostly what matters anyway, and why they have plenty of business. I’m not buying any more, especially since they cost the same to me.

1 Like
93

9 posts were split to a new topic: Graceful Exit after database removing

94

Maybe because i didn’t recreate all the databases manually?

96

Yes, but his db was corrupted in the RAID crash…

97

Were they? These were the first 2 posts by @joesmoe in this topic. He didn’t mention he had similar issues to the original poster.

100

Sorry, @BrightSilence, it escaped me that he is not the OP.

101

Cheers guys! Just wanted to pop in and confirm that I received the GE payment. It seems that not loosing hope did the trick :slight_smile:

11 Likes