Hard drive halts all transfers, resumes about 30 seconds later

[richdiesal]

I just built a new Ryzen machine on an Asus B350 board with a .5 TB SSD and 4 TB HDD. I've been experiencing random system hangs, where Google can't access its cache, I can't save, etc -
feels like the system stops responding but the mouse keeps working - and I've finally isolated the problem to the SSD, based upon incidents like this on in Resource Monitor (http://imgur.com/6HaZzzL):

I ran chkdsk - no errors, looks fine.
I swapped out the SATA cable for another, made no difference.

I am not sure what else to do to diagnose the cause, if this is a bad drive or something about the motherboard (which is admittedly new). The 4TB drive doesn't get accessed nearly as much, so I haven't seen the same problem with it yet.

 

[Rogue Leader]

Whats the brand and model on the SSD?

 

[richdiesal]

Whats the brand and model on the SSD?

It's this one: https://www.amazon.com/gp/product/B00TIJMEV2/ref=oh_aui_search_detailpage?ie=UTF8&psc=1
PNY 480GB CS1111

 

[Rogue Leader]

I would tend to believe its the drive hanging. I can't find much on that drive whether it has consistent issues, but IMO it looks like the drive may be overheating and the onboard controller locking up. That doesn't man that your system cooling is bad, more than likely its just the drive failing.

 

[Solandri]

Make sure you have all the motherboard drivers installed. I've seen a similar pause in disk access when I'd forgotten to install the chipset drivers. Go to the motherboard manufacturer's site, download all the drivers, and manually install them.

Win 7 it would flag a driver as missing in Device Manager. But Win 10 unhelpfully substitutes generic Microsoft drivers (which often don't work) when this happens and Device Manager shows no problems. So you're gonna have to install them all just to be sure.

 

[richdiesal]

I would tend to believe its the drive hanging. I can't find much on that drive whether it has consistent issues, but IMO it looks like the drive may be overheating and the onboard controller locking up. That doesn't man that your system cooling is bad, more than likely its just the drive failing.

The drive is brand-new - I also have now replicated this problem with a drive-to-drive transfer (from an SSD to HDD) and both cut off and on at the same time.

 

[richdiesal]

Make sure you have all the motherboard drivers installed. I've seen a similar pause in disk access when I'd forgotten to install the chipset drivers. Go to the motherboard manufacturer's site, download all the drivers, and manually install them.

Win 7 it would flag a driver as missing in Device Manager. But Win 10 unhelpfully substitutes generic Microsoft drivers (which often don't work) when this happens and Device Manager shows no problems. So you're gonna have to install them all just to be sure.

I did that but no luck! Perhaps this is a side effect of Ryzen? Perhaps a mobo problem that hasn't been patched yet?

 

[Rogue Leader]

Make sure you have all the motherboard drivers installed. I've seen a similar pause in disk access when I'd forgotten to install the chipset drivers. Go to the motherboard manufacturer's site, download all the drivers, and manually install them.

Win 7 it would flag a driver as missing in Device Manager. But Win 10 unhelpfully substitutes generic Microsoft drivers (which often don't work) when this happens and Device Manager shows no problems. So you're gonna have to install them all just to be sure.

I did that but no luck! Perhaps this is a side effect of Ryzen? Perhaps a mobo problem that hasn't been patched yet?

I doubt that, I have an 1800X on an Asus Crosshair VI and do not see issues like this.

In the transfer both cut at the same time, but the SSD could be causing the other to cut out.

What SATA ports are you plugged into, have you tried other ones?

 

[richdiesal]

Make sure you have all the motherboard drivers installed. I've seen a similar pause in disk access when I'd forgotten to install the chipset drivers. Go to the motherboard manufacturer's site, download all the drivers, and manually install them.

Win 7 it would flag a driver as missing in Device Manager. But Win 10 unhelpfully substitutes generic Microsoft drivers (which often don't work) when this happens and Device Manager shows no problems. So you're gonna have to install them all just to be sure.

I did that but no luck! Perhaps this is a side effect of Ryzen? Perhaps a mobo problem that hasn't been patched yet?

I doubt that, I have an 1800X on an Asus Crosshair VI and do not see issues like this.

In the transfer both cut at the same time, but the SSD could be causing the other to cut out.

What SATA ports are you plugged into, have you tried other ones?

The B350 board has 6 ports - I was in 6G_1. Just tried 6G_3, same problem.

 

[Rogue Leader]

I really feel like based on that its the drive thats bad out of the box.

 

[richdiesal]

I really feel like based on that its the drive thats bad out of the box.

Urgh, that is really not what I wanted to hear. Any hope that this is because it's on AHCI instead of RAID? I have read in a few places that is sometimes an issue with SSDs, but I don't want to reinstall Windows unnecessarily.

 

[Rogue Leader]

I really feel like based on that its the drive thats bad out of the box.

Urgh, that is really not what I wanted to hear. Any hope that this is because it's on AHCI instead of RAID? I have read in a few places that is sometimes an issue with SSDs, but I don't want to reinstall Windows unnecessarily.

Definitely not. You would only use RAID if you need it specifically (Multiple drives linked - mirrored, striped, etc).

 

[richdiesal]

I really feel like based on that its the drive thats bad out of the box.

Urgh, that is really not what I wanted to hear. Any hope that this is because it's on AHCI instead of RAID? I have read in a few places that is sometimes an issue with SSDs, but I don't want to reinstall Windows unnecessarily.

Definitely not. You would only use RAID if you need it specifically (Multiple drives linked - mirrored, striped, etc).

Ok, drat. New symptom though, now that I'm using the system more: throughput on both drives is quite poor. On both drives, it takes about 7 seconds to create a folder, consistently - the system hangs after I press enter after giving it a new name. Large-scale file transfer rates are also abysmal - in a system-to-system transfer over homegroup on 802.11n, I was getting about 100kBps max. This feels like it's pointing toward the mobo?

I've also now seen two "memory management" BSODs, but my suspicion is that problem is caused by the same thing causing these other issues.

 

[Rogue Leader]

I mean its a stretch but it could be some failure in the motherboard. I would still think its the SSD first. Then yes its possible the SATA controller is failing.

Do you have another system you could test the SSD in.

 

[richdiesal]

I mean its a stretch but it could be some failure in the motherboard. I would still think its the SSD first. Then yes its possible the SATA controller is failing.

Do you have another system you could test the SSD in.

A problem with the SSD could cause slowdowns on the HDD?

I unfortunately don't have access to another system without buying a SSD enclosure from somewhere, and my return window on the SSD is shrinking.

 

[Rogue Leader]

I mean its a stretch but it could be some failure in the motherboard. I would still think its the SSD first. Then yes its possible the SATA controller is failing.

Do you have another system you could test the SSD in.

A problem with the SSD could cause slowdowns on the HDD?

I unfortunately don't have access to another system without buying a SSD enclosure from somewhere, and my return window on the SSD is shrinking.

Yeah if its making the sata controller wait somehow. Also you said the HDD problems are when transferring between the two drives right?

 

[richdiesal]

I mean its a stretch but it could be some failure in the motherboard. I would still think its the SSD first. Then yes its possible the SATA controller is failing.

Do you have another system you could test the SSD in.

A problem with the SSD could cause slowdowns on the HDD?

I unfortunately don't have access to another system without buying a SSD enclosure from somewhere, and my return window on the SSD is shrinking.

Yeah if its making the sata controller wait somehow. Also you said the HDD problems are when transferring between the two drives right?

That's how I initially spotted the problem, but I am also having delays on the HDD on its own. I can actually copy files off my 802.11n network to either drive at the same speed currently (although both seem slow, about 4 MBps). I'm wondering if the initial speed issues were because I encountered a drop to zero and then Windows did not recover well.

 

[Rogue Leader]

There could be some corruption somewhere but I doubt it. I would still start with the SSD and work from there. A bad component can cause all kinds of weird issues.

 

[richdiesal]

There could be some corruption somewhere but I doubt it. I would still start with the SSD and work from there. A bad component can cause all kinds of weird issues.

Ok, situation resolved, I think. I learned I was apparently misinterpreting the resource monitor graph - the green lines are queued read/write events and the blue line is drive activity - so it was 100% drive activity, not 0%.

After reading up on 100% disk utilization on SSDs, I made two changes that fixed it which I am adding here in case anyone comes across this with the same problem I had:
1) Disable MSI mode: https://support.microsoft.com/en-us/help/3083595/task-manager-might-show-100-disk-utilization-on-windows-10-devices-with-message-signaled-interrupt-msi-mode-enabled
2) Disable Superfetch (option 2 on this page): https://www.drivereasy.com/knowledge/fix-100-disk-usage-in-task-manager-improve-pc-performance-on-windows-10/

Since these worked (and I seem to need both to completely eliminate the problem), my suspicion is that Windows 10 was not correctly identifying my SSD as an SSD, which should disable both of these anyway but just for that drive. That may be why I found information that some people switching from AHCI to RAID could fix this - MSI mode seems to be a feature enabled by AHCI, so disabling AHCI might fix that issue too. Alternatively, switching to RAID might force the drivers to reload or for Windows to recheck that the drive is an SSD. But in either case, that is a "baby with the bathwater" sort of fix - disabling these two features alone is enough.

Before changing these two things, I would hit 100% for 20 seconds or so at least once every 10 minutes, although it never happened while gaming - only while doing "productivity" sorts of tasks (using statistical software, Word, etc - lots of simultaneous applications). I have now not had a slowdown or crash in about 6 hours of continuous usage since those changes, so I am hopeful this will continue!

 

[Rogue Leader]

Thats very strange, I've built many computers with SSDs in AHCI mode and never seen an issue before. While 100% useage does look like the graph you showed me how you described it was that it was locking up, so I assumed a freeze. I didn't realize it was still showing data transferring during that time period.

 

[richdiesal]

Thats very strange, I've built many computers with SSDs in AHCI mode and never seen an issue before. While 100% useage does look like the graph you showed me how you described it was that it was locking up, so I assumed a freeze. I didn't realize it was still showing data transferring during that time period.

Yeah, that's just how I was experiencing it - the mouse would work, and I could still use the system as long as it didn't need to access the hard drive. Any time it did, e.g., if Chrome tried to access its cache, Chrome would hang. But it would recover after 20-30 seconds as if nothing had happened. What was weird is that there was no data usage spike when it happened - even though it went to 100%, no new read/writes were appearing - all the Bytes/sec readings stayed about the same as pre-hang.

I am wondering if this could be a side effect of going cheap on the SSD (PNY). I discovered recently that PNY is not on the sizable list of "good vendors" of SATA devices for the Ryzen/ASUS board.

 

[Rogue Leader]

Could be some sort of drive caching issue, due to it being such a cheap drive.