longer than normal bootup times from power off - cause for concern for NVMe drive?

[philb2]

The drive is a Sabrent Rocket 1TB PCIE-4, purchased around April, 2020 with 6800 total power on hours, per HD Sentinel Pro. It started out life in an ASUS X570e motherboard, then was moved over to an ASUS X670e motherboard (with no real issues).

Sometimes when I turn on the system from power off, I hear the "speaker ping" sound pretty quickly, but then it takes longer than normal for me to get to the multi-boot menu. From that point on, boot process is normal. Stable operating temp is 45 C, health is 100%. I'm no expert on SMART parameters, but it all looks OK to me.

Once a Muskkin SSD drive failed on me without warning, and I certainly don't want this drive to fail like that. What should I look out for? Is there a case for replacing the drive based on age or power-on hours? I would probably replace this drive with a 4TB drive, so I can replace one of the spinning rust drives in my system.

 

[Dopamin3]

It sounds like you're talking about your PC's POST time, which realistically doesn't have anything to do with your NVME.

SSDs can fail without warning, so there isn't a reliable way to predict this. SMART data, and programs like HD Sentinel Pro or CrystalDiskInfo that give you a health percentage can help. But what usually happens, and what probably would've happened to your Mushkin SSD it just died without providing any critical SMART values beforehand.

The manufacturers give SSDs a MTBF which is usually ridiculous like 1 or 2.5 million hours, and they give a TBW saying it should be able to handle that many terabytes of data written. You can easily exceed the rated TBW and the drive keeps kicking, conversely it could fail long before hitting the TBW. Just always have a backup of cricitical data.

 

[dogDAbone]

Well, sorry to add to your misery, I can tell you that the last rig I worked on a few weeks ago, had the same drive in it, and was brought to me because it suddenly failed to boot, seemingly for no reason.....

I went thru a few hours of diagnostics using various apps & test processes, after booting the machine from another drive that I keep just for these situations....but no joy there either

Fortunately the client had a recent back up of all their apps & important files, so I wiped the drive clean using multi-pass zeroing & then copied some random small files to it, which appeared to complete successfully, but the files could not be found anywhere.

Even after all that, the drive would not show up in disk mgmt or my partitioning apps to be initialized....so I chucked it & sold the person a new 850X, upon which I installed W10Pro and used to restore everything from his back-up....and since then, nottaproblemo whatsoever

 

[pendragon1]

sounds like memory training, maybe

 

[philb2]

Well, sorry to add to your misery, I can tell you that the last rig I worked on a few weeks ago, had the same drive in it, and was brought to me because it suddenly failed to boot, seemingly for no reason.....

I went thru a few hours of diagnostics using various apps & test processes, after booting the machine from another drive that I keep just for these situations....but no joy there either

Fortunately the client had a recent back up of all their apps & important files, so I wiped the drive clean using multi-pass zeroing & then copied some random small files to it, which appeared to complete successfully, but the files could not be found anywhere.

Even after all that, the drive would not show up in disk mgmt or my partitioning apps to be initialized....so I chucked it & sold the person a new 850X, upon which I installed W10Pro and used to restore everything from his back-up....and since then, nottaproblemo whatsoever

Yes you are adding to my misery. Fortunately I already have a robust backup routine that runs daily.

 

[philb2]

sounds like memory training, maybe

OK, but why would my system still be doing memory training, since I built my rig well over a year ago. My system is rock-steady, so I haven't updated

It sounds like you're talking about your PC's POST time, which realistically doesn't have anything to do with your NVME.

SSDs can fail without warning, so there isn't a reliable way to predict this. SMART data, and programs like HD Sentinel Pro or CrystalDiskInfo that give you a health percentage can help. But what usually happens, and what probably would've happened to your Mushkin SSD it just died without providing any critical SMART values beforehand.

The manufacturers give SSDs a MTBF which is usually ridiculous like 1 or 2.5 million hours, and they give a TBW saying it should be able to handle that many terabytes of data written. You can easily exceed the rated TBW and the drive keeps kicking, conversely it could fail long before hitting the TBW. Just always have a backup of cricitical data.

Have not updated the BIOS or the chipset driver in a long time. No BSODs. No overclock except BPO.

I know I should update the BIOS, the chipset drivers and try manual O/C, but I have so many other things that I need to do which use my PC.

 

[philb2]

It sounds like you're talking about your PC's POST time

Yes.

, which realistically doesn't have anything to do with your NVME.

Good to know. One less thing to add to my "worries" list.

SSDs can fail without warning, so there isn't a reliable way to predict this.

I wish that someone can come up with a diagnostic that would probe the innars of an NVMe drive.

The manufacturers give SSDs a MTBF which is usually ridiculous like 1 or 2.5 million hours, and they give a TBW saying it should be able to handle that many terabytes of data written

I guess that unlke HDDs, there isn't anyone like Backblaze to provide performance stats on a large installed base of drives.

 

[pendragon1]

OK, but why would my system still be doing memory training, since I built my rig well over a year ago. My system is rock-steady, so I haven't updated

idk, maybe the battery is getting weak. maybe you havent set it to retain the memory training and its off for long enough that it wants to redo it? do you have any other drives connected? maybe something else is slowing down posting? all we know is the nvme and x670e board, but once it post its boot normal. so i wouldnt be suspecting the drive first...

 

[philb2]

idk, maybe the battery is getting weak. maybe you havent set it to retain the memory training

I'll need to check the BIOS. No time right now. Too much else that HAS to be done today. (ain't life grand )

and its off for long enough that it wants to redo it?

8-10 hours overnight?

do you have any other drives connected?

Just some SATA SSDs and HDDs.

maybe something else is slowing down posting? all we know is the nvme and x670e board, but once it post its boot normal. so i wouldnt be suspecting the drive first...

OK. I do intend to upgrade that NVMe drive, but I'm hoping I can outwait the recent price increases.

 

[pendragon1]

I'll need to check the BIOS. No time right now. Too much else that HAS to be done today. (ain't life grand )

8-10 hoiea overnight?



Just some SATA SSDs and HDDs.

OK. I do intend to upgrade that NVMe drive, but I'm hoping I can outwait the recent price increases.

yup.
yup.
try disconnecting those.
you might have plenty of time. grab the oem's ssd tool and have it do a health check.

 

[Dopamin3]

I wish that someone can come up with a diagnostic that would probe the innars of an NVMe drive.

You already get that with SMART data reported, and this goes for hard drives too. But there is no such thing as a crystal ball, so it can't predict sudden failures. But often times you can pre-emptively replace it when things are going south.

Also when you get new storage devices, doing a full pass of badblocks isn't a bad idea either. Storage mediums that report healthy SMART can sometimes have unusable blocks and doing a whole pass saves you hassle in the long run. Badblocks can write and read to the entire space of the drive and let you know if it's okay or not.