Toshiba MAMR experience?

Keep in mind that the MG series is an enterprise class HDD. They expect you to have proper cooling like in a Datacenter to cool your HDDs. I think this is the reason they set the max temperature so low

1 Like

Just received a 20TB return for my latest 18TB RMA. Can’t wait to have them all replaced. :sunglasses:

2 Likes

Didn’t that exchange cost you a node, when the 18TB died?

I am replacing the disks when I see first pending or uncorrectable sectors. So far no node was lost, some missing piece files out of millions don’t matter.

Ok, then I understand you are happy with the exchange :slightly_smiling_face:

Until now 21 of my TOSHIBA MG09ACA18TE have failed while other brands and other Toshiba models have zero fails under same conditions.

2 Likes

OMG! That’s huge! How many do you have? Are they covered by warranty?

I have several MG9 or mg10 18tb none of them failed, stable cooling is very needed.

I also use several of these and none have failed so far. Good cooling is provided so that the hard drives do not exceed 40°C.

1 Like

Is this what specified in the data sheet? Most disks operating temperature is up to 60C. If these specific disks fail when ran within specs — they are poorly designed.

Why do people keep buying that junk?

I too had never have Toshiba drive not fail on me. At this point it’s a systemic issue. Just stop buying them.

And no, I’m not going to lower the temps. My disks run at 55-60C.

1 Like

I had unusually high failure rate with MG08 16TB CMR models in JBOD enclosures with plenty of airflow.
They usually started to fail between two and three years in, in a way they were unable to read random sectors, but not always showed this in SMART.
I still run a bunch of them and dmesg is just a mess because of them.

Funny enough, just couple of days ago I was looking at used drives market and one of the sellers was selling various WDs, Seagates and Toshibas.
All were used, WDs and Seagates were listed as zero bad sectors, the Toshibas on the other hand were all listed as “some bad sectors”.

Anyways, better to buy used (I haven’t bought new for years now and couldn’t be happier) and probably CMR.
All these MAMR, HAMR might still be problematic I think.
And beware, because some of the recertified Seagates for example, in capacities normally using CMR/PMR, are actually derated HAMR drives that did not pass factory tests as full capacity drives. And these apparently are also spec derated - like lower read and write speeds etc.
Similar bullshit like WD did with CMR drives that were actually SMR drives.

I also avoid HAMR drives. Do you have a source for what you say about the risk of receiving a factory recertified drive with a lower capacity than the drive’s original capacity? That’s new to me. Maybe examples of this are limited to third-party refurbishers?

This discussion here for example: Seagate Exos Recertified Hard Drives Buying advice / Awareness. : r/DataHoarder.

I saw examples of the mentioned drives at one of the refurbed drives reseller sites, and because of this, instead of buying Seagates that looked like CMR (and in such capacities and in that line Seagate is manufacturing them as CMR without assisting), I opted for WD, which for capacities where Seagate is still using CMR, WD is assisting, but with current (stronger/more pointy magnetic field), no heating or other radiation is being used. It basically is still the same platters tech in the WD, so reliability is well known.
The price at the time was about the same as Seagate, but Seagates were this binned down drives.
But at capacities larger than 24TB I think all of them will be some sort of HAMR/MAMR tech, so generally it is probably better to avoid larger capacities at this time.

1 Like

Ok, that is good to know, thanks.

It seems like backblaze now facing similar problems with Toshiba…

If you see exactly problem is only with this additional function
MG08ACA16TEY
SIE: Sanitize Instant Erase. SIE is a function to invalidate the data recorded on the magnetic disks at a blink.

this Y in end, means SIE, looks like it kill disk faster.

Doesn’t it say the drives aren’t unreliable: that they were only taken offline to have firmware updated (for additional performance)?

Doesn’t sound like performance optimization. Unfortunately Toshiba is not offering firmware updates to normal users.

In the same article… “working with Toshiba, we deployed some firmware updates they provided to optimize performance on these drives. Because we needed to pull drives to achieve this in some cases, we had an abnormal number of “failed” drives in this population.

It’s unfortunate if they’re not providing firmware improvements to everyone though :frowning:

No firmware for the masses means they don’t give a F about us, so we shouldn’t buy them in the first place. Seagate has even 6 updates for a model, publicly available. Thumbs up for them.