HDD Massive slowdowns during file copy and installs

[bobrooney]

Recently I've been having massive slowdowns when copying files after a months use of windows 8.1. I have 8 drives.

4x1tb sata2 WD green
1x500gb sata2 WD 7200rpm black drive
2xhitachi 7200rpm 2tb drives,
1xsandisk extreme pro ssd - main drive

i manly copy between the two hitachi drives and really thought that they were going bad until i tested copy functions on all the drives with the same result. no way 8 drives are going bad at the same time! i normally hit 100-125MB/s copy speeds on 1GB files. now half way through the copy, the speed just drops to 2-4MB/s and is continuous with that speed if I immediately copy another file. If I wait a minute or two and start copying again, 100-125MB/s copy speeds and then the drop again. its like the cache isnt working right or something.

Thought I'd see if someone had this issue with some solution before upgrading back to windows 7 ultimate. the drives are under 2 controllers on my gigabyte 870A-UD3 motherboard. 2 green drives are set under IDE mode under one controller that operates under sata2, and the rest are set as AHCI sata3. I have no idea what might be causing this. there are no 100% disk usage issues showing in the task manager. i even disabled indexing, prefetching, and superfetching. Any ideas?

 

[Androidcity]

I have the same thing, driving me crazy. Its slows down my whole system

 

[altae]

Why do you run the HDDs in different modes (IDE and AHCI)? Using a modern os like Windows 8.1 you are supposed to run the hdds in AHCI mode (or raid mode if you use a raid setup). Get that right first and then report back.

 

[Cliff S]

Are all your disks policies set up like this(best when/if you have ,an uninterruptible power supply (UPS) to provide a backup power source for your computer too guard against power failures) with buffer flushing on the system has to wait until the buffer is written then it can fill with more data(this is safer) turned off the buffer can fill and write at the systems convenience and to the user it looks as if it's already written. Read this

 

[crawfish]

The top checkbox is safe with UPS or laptop battery, the bottom one apparently not so much, as the drive itself is supposed to have its own separate PSU. This is the article that needs to be read concerning that option:

Dangerous setting is dangerous: This is why you shouldn't turn off write cache buffer flushing - The Old New Thing - Site Home - MSDN Blogs

In any case, enabling it is not necessary to avoid the problem described by the OP. When I copy large files between HDs, and the process degrades to the actual speed of the drives, it's not 2-4 MB/sec. It's more like 40-100 MB/sec depending on the drives involved. This is true even for the backup drives I've used in eSATA and now USB3 docks.

OP, are you sure you're not copying within the same drive, perhaps between partitions hosted on the same drive? Is there any chance you're launching multiple copy operations at once involving the same drive(s)? I would guess probably not since you've thought of other things that might try to concurrently access the drives, though the features you've cited are supposed to get out of the way on their own.

 

[bobrooney]

Why do you run the HDDs in different modes (IDE and AHCI)? Using a modern os like Windows 8.1 you are supposed to run the hdds in AHCI mode (or raid mode if you use a raid setup). Get that right first and then report back.

no raid at all. two of those 1TB sata2 WD green drives are encrypted, and unfortunately, take half the write performance hit under AHCI. When I initially enabled them under IDE mode, their copy performance was normal. then over a months time the copy issue showed up, as with all the drives.

 

[bobrooney]

Are all your disks policies set up like this(best when/if you have ,an uninterruptible power supply (UPS) to provide a backup power source for your computer too guard against power failures) with buffer flushing on the system has to wait until the buffer is written then it can fill with more data(this is safer) turned off the buffer can fill and write at the systems convenience and to the user it looks as if it's already written. Read this

I dont have the "Turn off Windows write-cache buffer flushing" checked off any of my drives. Never had to do that before. Going to give it a try and see what happens.

 

[Cliff S]

Are all your disks policies set up like this(best when/if you have ,an uninterruptible power supply (UPS) to provide a backup power source for your computer too guard against power failures) with buffer flushing on the system has to wait until the buffer is written then it can fill with more data(this is safer) turned off the buffer can fill and write at the systems convenience and to the user it looks as if it's already written. Read this

I dont have the "Turn off Windows write-cache buffer flushing" checked off any of my drives. Never had to do that before. Going to give it a try and see what happens.

I have mine set this way, but I have a very good power source(still taking chances & I use Macrium Reflect). Where I live Germany I haven't experienced any unannounced power outages(from the city) or ones caused from storms. Its a lot different in N.H. where I come from.

 

[RolandJS]

Not likely here, however, the combination of multible read/write operations along with assertively set security programs can produce such slowdowns.

 

[bobrooney]

one thing i did not mention was the read speeds were not affected when copying to usb 3.0 external hard drives that reached 150Mb/s in write speeds when copied to.

 

[bobrooney]

The top checkbox is safe with UPS or laptop battery, the bottom one apparently not so much, as the drive itself is supposed to have its own separate PSU. This is the article that needs to be read concerning that option:

Dangerous setting is dangerous: This is why you shouldn't turn off write cache buffer flushing - The Old New Thing - Site Home - MSDN Blogs

In any case, enabling it is not necessary to avoid the problem described by the OP. When I copy large files between HDs, and the process degrades to the actual speed of the drives, it's not 2-4 MB/sec. It's more like 40-100 MB/sec depending on the drives involved. This is true even for the backup drives I've used in eSATA and now USB3 docks.

OP, are you sure you're not copying within the same drive, perhaps between partitions hosted on the same drive? Is there any chance you're launching multiple copy operations at once involving the same drive(s)? I would guess probably not since you've thought of other things that might try to concurrently access the drives, though the features you've cited are supposed to get out of the way on their own.

even if i am copying to the same drive, i should not be hitting 2-4Mb/s. interestingly enough, i just tried to copy a file to the same drive and got decent results, around 50Mb/s with no drops. The one last thing I might try is to replace the sata cables.

 

[jpolacko]

This is how I have my drive set up under device management! I don't experience any drive performance issues.

 

[klbterminator]

- how long is it since last time you defrag? that's might be a problem because windows do stop writing to RAM at some points when copying
- try to turn off the Real-time protection in Windows Defender and test the drive, it constantly checks the IO (had hard time playing LoL with it)

 

[PlatypusKnight]

Was this issue ever solved?

I have a similar issue. jpolacko and I have similar systems. Why slow down on one and not on another. I also note he runs Norton and I run MS Defender, a much less resource intensive AV...from what I understand.

 

[bobrooney]

Was this issue ever solved?

I have a similar issue. jpolacko and I have similar systems. Why slow down on one and not on another. I also note he runs Norton and I run MS Defender, a much less resource intensive AV...from what I understand.

performance did return once i figured out which of the drives were giving me issues and changed the sata cables. AHCI and IDE modes have been left unchanged. i am considering moving the 4 green drives into a 4-bay enclosure and running them off the e-sata or usb 3.0 from the back of my motherboard. maybe the motherboard was not meant to have all sata ports connected at the same time.

 

[PlatypusKnight]

[
performance did return once i figured out which of the drives were giving me issues and changed the sata cables

So it was the cables?

That's not something I wanted to hear.

 

[bobrooney]

[
performance did return once i figured out which of the drives were giving me issues and changed the sata cables

So it was the cables?

That's not something I wanted to hear.

it could have also been the mainboard. i didnt keep tabs on what ports the drives were plugged into. its one of the reasons why i need to get a a 4-bay external enclosure or at least a 4-bay drive dock. i have a feeling that the old cables still might be good, and that the ports are acting up once in a while. maybe the old cables needed to be re-seated, plugged out and plugged back in. i dont know...i really didnt have much time to speculate.