Possible Failing Disk?

[xCatalystx]

The power "issue" is a non starter.

I should have made myself more clear. Less so on drive usage more so on cooling. We're I live it gets pretty warm (even in winter) and very hot during summer so I tend to aim more for lower power, cooler running drives. I do take steps to keep the system cooled but producing less heat makes it a bit easier as the system is fairly compact.

I know when I was running the 7200rpm HST drives I actually struggled to keep them cool during certain times of the year without additional cooling (aircon) when not idle. Unfortunatly with where I live and WAF (wife approval factor), there are only so many places I can keep the equipment (inc homelab).

But that is kind of outside the scope.

I can quote a lot of numbers to show you why the Gold is better than Red NAS, but it would be better for you to look at the numbers a little bit.

Yeah, I've been reading a fair amount of data the last month to make some choices.

The real difference is going to be price. How much extra are you willing to pay for an Enterprise drive with better specs, warranty, durability, and performance?

Problem is unless there is a great deal on Amazon Global, AU prices can be pretty savage for enterprise gear. Hear's hoping Black Friday has some good deals.

I was kind of wondering if it would be safe to use 4x10tb drives in raidz2 or maybe 5x10tb raidz3. Buying fewer drives (4-5) would have a few advantages in my case. I think it would be fine as i keep an onsite backup, but once again weighting options =)

BLACK FRIDAY HURRY UP!!!

using WD Gold since you like them

not brand bias, but work liked the wd re's than went to gold. I've been happy with almost anything (WD, Toshiba, HGST) but Seagate which seems to die on me like nothing.

Anyways thanks agian.

 

[joeschmuck]

@xCatalystx have you looked at my Hard Drive Troubleshooting Guide? See my signature.

First of all, your drive may be showing you that it is in fact failing however it may be fine for a while longer. In other words I see no urgency to replace it immediately.

So in looking at your SMART data the valude for ID1 is of consern becasue this WD drive normally only reports true read errors and you do have some read errors but this is not fatal. Your Extended Test also shows that you have read errors as well as ID 197 Pending Sectors which is a value of 8. The ID 200 Multi-Zone errors will likely clear up on thier own once the offending sector(s) have been locked out.

After your Extended test finishes up, post the output of smartctl -a /dev/da4 or if you want use the -x switch instead. Also perform a Scrub and wait for it to finish then check the status to ensure all is good. I suspect the scrub results will be good.

As for your RMA, if the website says it's expired, odds are against WD performing a RMA for it but you can try to contact them and tell them the failure occured a few day earlier and see if they will accept it, but don't hold your breath.

Now comes the last part of troubleshooting your drive, this is something I recently performed myself and it resulted in locking out a single sector and my drive has been working perfectly for the past several months, not even a small complaint. The trick is you will take this drive out of service for the badblocks testing. In my guide I use the failing LBA and focus on heavy testing the small portion of the drive where this LBA failure occured. Since you are already looking for a replacement drive then you might want to wait to do this type of testing but I leave that up to you.

ALSO if you find 6TB drives on sale, give me a shout out! I too am looking for four 6TB hard drives and with all the sales going on, I'm looking hard but I want them cheap. I'll take 8TB drives too if they are cheap. I'm looking at the $160 or less price range. If I can't find them cheap then I'm waiting until I have a real drive failure and I'm forced to buy more drives.

 

[farmerpling2]

I should have made myself more clear. Less so on drive usage more so on cooling. We're I live it gets pretty warm (even in winter) and very hot during summer so I tend to aim more for lower power, cooler running drives. I do take steps to keep the system cooled but producing less heat makes it a bit easier as the system is fairly compact.

I know when I was running the 7200rpm HST drives I actually struggled to keep them cool during certain times of the year without additional cooling (aircon) when not idle. Unfortunatly with where I live and WAF (wife approval factor), there are only so many places I can keep the equipment (inc homelab).

The Helium 7200 RPM drives are cooler, but probably not as much as a 5x00 RPM drive.

Over the years, I have used this little saying, "If the watts being used by the drive are roughly the same, then the heat created by the drives will not be much different with similar physical makeup. (i.e. 3.5", 4 platters, etc. " Enterprise drives usually have heavier drives then consumer/commercial - the metal portions have more metal.

See if you can "borrow" the WD Gold and add it to your setup. Can you get a infrared thermometer? If so, you can then test the heat of the drive. There is a certain point that the drive manufacturers use to measure the heat, look at drive manual PDF for more information on this.

Generally, the older HGST typically run hotter and use more watts. The newer ones are pretty good.

not brand bias, but work liked the wd re's than went to gold. I've been happy with almost anything (WD, Toshiba, HGST) but Seagate which seems to die on me like nothing.

I run HGST enterprise because they just work and under a heavy workload they still work. That does not mean the other drives are bad. I would get WD Gold without a thought. Seagate enterprise would be fine, also. Toshiba would be my last choice and I would likely not use them. Just my personal experience and opinion.

When I use Seagate consumer drives, I run them in windows software RAID1
configuration. There has always been a failure of one of the members that needs to be RMA'd- again consumer, not enterprise.

Regarding failing drive, I would wipe it a couple more times. I have seen it take 2-3 runs before the errors stop occurring.

joeschmuck has a good idea about focused testing. Keep in mind that the failing magnetic area occurs in multiple dimensions. You want to test an area 360 degrees around the known failing LBA. You may clean up a sector but find out later an adjacent sector starts failing.

I have had similiar failures like you are having and wiped the drive 3 or 4 times and the drive is still working without a problem 1+ years later.

You also mention about using fewer disks that are larger to fit your finances. Know your workload, as fewer disks means slower transfer rate. A rebuild or scrub will take longer also, possibly a lot longer if you have your drive space usage 50% or 60%+.

 

[farmerpling2]

ALSO if you find 6TB drives on sale, give me a shout out! I too am looking for four 6TB hard drives and with all the sales going on, I'm looking hard but I want them cheap. I'll take 8TB drives too if they are cheap. I'm looking at the $160 or less price range. If I can't find them cheap then I'm waiting until I have a real drive failure and I'm forced to buy more drives.

Maybe we should create a thread for this. I can always let people know when I see something. We have a couple people who watch ebay like a hawk who I am sure would drop a note.

 

[joeschmuck]

Maybe we should create a thread for this. I can always let people know when I see something. We have a couple people who watch ebay like a hawk who I am sure would drop a note.

This would be a nice Off-Topic thread to have. NewEgg has a sale on HGST 4TB drives (four pack) for $400 which is a very fair price, I even considered buying them however I'd need to buy one more drive to get the capacity I desire in a RAIDZ2 configuration. I'm still holding out for the 6TB drives as that is what I really prefer. Of course I could drop down to a RAIDZ1 with the 4TB drives but I'd rather not.

joeschmuck has a good idea about focused testing. Keep in mind that the failing magnetic area occurs in multiple dimensions. You want to test an area 360 degrees around the known failing LBA. You may clean up a sector but find out later an adjacent sector starts failing.

When I tested my drive out for the failing sector I started at +/- 10,000 of the failing LBA with Badblocks and after I beat the crap out of it there and all worked fine I then increased it to +/- 100,000 of the failing LBA and I think I even went as far as +/- 200,000 LBAs and then the entire drive. This drive is where my data is backed up and I want it safe so if testing takes me a few days to be sure the drive is not going to just keep failing, then that is what I'll do. @farmerpling2 is correct, there can be and likely will be adjacent sector damage. Typically when a sector fails it is because the magnetic surface is coming off of the platter and when this starts to peal off it may start with one small section of the platter but it can and typically will grow over time. If you are lucky then you will be able to test your drive for only a few days before placing it back in service. The guide I created tells you how to run this testing if this is the path you want to take. Even if you don't plan to use the drive I think it would be nice to test the drive and if the results are favorable then place it into a system and see how long it lasts.

 

[xCatalystx]

Sorry for the late reply. I swear, i never get sick in winter, but the change to warm weather just destroys my body >_>.

See if you can "borrow" the WD Gold and add it to your setup.

Yep, I've borrowed two actually (just incase). Working great so far, no errors in the pool after the scrub and extended scan on the disks (all). so safe for now.

WD got all weird about the warranty but have agreed to RMA it as i presented proof of purchase that contradicts there warranty date. I was planning on reusing the 3tb drives later anyways as another backup offsite so that's a win.

Generally, the older HGST typically run hotter and use more watts. The newer ones are pretty good.

yep that's what I have seen aswell.

In my guide I use the failing LBA and focus on heavy testing the small portion of the drive where this LBA failure occured.

I was planning on doing this if the drive was not going to be covered by RMA, but now I am less concerned. Added it to my wiki/notes for future reference.

You also mention about using fewer disks that are larger to fit your finances. Know your workload, as fewer disks means slower transfer rate. A rebuild or scrub will take longer also, possibly a lot longer if you have your drive space usage 50% or 60%+.

This is the thing that is bothering me atm. Speed is less of an issue considering I am happy to just saturated 1gbe for reads. My other pool running SSD's for my VMS / testing / iscsi quite happy pushing multiple 1gbe. I'll work it out, a lot will depend on what sizes are sale.

Also, Off-Topic, might also be worth keeping an eye out on https://www.reddit.com/r/DataHoarder/ they tend to post deals aswell for this sort of thing =P