Hello

 

I recently built an unraid server and noticed that my write speeds have been decreasing. Here is what i started it off with when i was testing it out.

 

Unraid version 5.0.3

Asus H87I Plus with Intel i3 4330

4GB of GSKILL DDR31600

Parity - Seagate SATA III - ST3000DM001

Disk1 - Seagate SATA III - ST3000DM001

 

When I started this, i was getting transfer speeds between 60-80MB

Since then, I have added the following:

 

Disk 2 - SATA II - WDC WD20EARS-00MVWB0

Disk 3 - SATA II - WDC WD20EARS-00MVWB0

 

Now when I copy things over, it seems my write speeds have dropped anywhere from 20MB - 40MB.

When i look at the sys logs i see the following

 

Dec 25 14:57:35 UnRAID kernel: ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300) (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata4: SATA link up 3.0 Gbps (SStatus 123 SControl 300) (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata2: SATA link up 6.0 Gbps (SStatus 133 SControl 300) (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300) (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata5: SATA link up 6.0 Gbps (SStatus 133 SControl 300) (Drive related)

 

To me is looks like it recognizes the attached drive speeds but Then I see this

 

Dec 25 14:57:35 UnRAID kernel: ata1.00: ATA-9: ST3000DM001-1CH166,            W1F2WY46, CC27, max UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata1.00: 5860533168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata2.00: ATA-8: ST3000DM001-1CH166,            Z1F13JPP, CC43, max UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata2.00: 5860533168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata2.00: configured for UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ACPI Error: [DSSP] Namespace lookup failure, AE_NOT_FOUND (20130117/psargs-359) (Errors)

Dec 25 14:57:35 UnRAID kernel: ACPI Error: Method parse/execution failed [\_SB_.PCI0.SAT0.SPT2._GTF] (Node f307b540), AE_NOT_FOUND (20130117/psparse-537) (Errors)

Dec 25 14:57:35 UnRAID kernel: ata4.00: ATA-8: WDC WD20EARS-00MVWB0,      WD-WCAZA2696321, 51.0AB51, max UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata4.00: 3907029168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata3.00: ATA-8: WDC WD20EARS-00MVWB0,      WD-WCAZA3005844, 51.0AB51, max UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata3.00: 3907029168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA (Drive related)

Dec 25 14:57:35 UnRAID kernel: ACPI Error: [DSSP] Namespace lookup failure, AE_NOT_FOUND (20130117/psargs-359) (Errors)

Dec 25 14:57:35 UnRAID kernel: ACPI Error: Method parse/execution failed [\_SB_.PCI0.SAT0.SPT2._GTF] (Node f307b540), AE_NOT_FOUND (20130117/psparse-537) (Errors)

Dec 25 14:57:35 UnRAID kernel: ata4.00: configured for UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata3.00: configured for UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ACPI Error: [DSSP] Namespace lookup failure, AE_NOT_FOUND (20130117/psargs-359) (Errors)

Dec 25 14:57:35 UnRAID kernel: ACPI Error: Method parse/execution failed [\_SB_.PCI0.SAT0.SPT4._GTF] (Node f307b5d0), AE_NOT_FOUND (20130117/psparse-537) (Errors)

Dec 25 14:57:35 UnRAID kernel: ata5.00: ATA-8: MKNSSDCR120GB, MK130927AS1120395, 521ABBF0, max UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata5.00: 234441648 sectors, multi 16: LBA48 NCQ (depth 31/32), AA (Drive related)

Dec 25 14:57:35 UnRAID kernel: ACPI Error: [DSSP] Namespace lookup failure, AE_NOT_FOUND (20130117/psargs-359) (Errors)

Dec 25 14:57:35 UnRAID kernel: ACPI Error: Method parse/execution failed [\_SB_.PCI0.SAT0.SPT4._GTF] (Node f307b5d0), AE_NOT_FOUND (20130117/psparse-537) (Errors)

Dec 25 14:57:35 UnRAID kernel: ata5.00: configured for UDMA/133 (Drive related)

Dec 25 14:57:35 UnRAID kernel: ata1.00: configured for UDMA/133 (Drive related)

 

 

Looking at this log line "configured for UDMA/133 (Drive related)"does this mean its not using SATA III speeds? Also i see some errors but not sure what they mean.

 

As a test, I've added a cache drive (SATA III 120GB SSD MKNSSDCR120GB) and see write speeds now anywhere from 60MB - 80MB but i'm hoping to get these speeds without a cache drive as my original setup seemed to perform pretty well.

 

Any info would be greatly appreciated.

 

Thanks in Advance

 

 

 

 

 

 

 

 

 

 

 

 

Looks like maybe my parity was being built when I got those original speeds.

Just turned the parity off and tested a few file transfers and speeds where between 70-90MB.

 

If that is the case then it looks like there is no way around it

 

 

 

When you write to an unprotected (no parity) array, you're writing directly to the disk.    When you write to a protected array, the system has to do FOUR disk operations -- two reads (the target disk and parity disk); followed by two writes (to the target disk and the parity disk).    This is necessary to ensure parity is always up-to-date.

 

There's no way to circumvent that requirement; but you CAN cause the system to defer the writes to the protected array by using a cache drive.  Then the writes you do to a cache-enabled share will actually be to the cache drive ... and they'll later be moved to the protected array at a time when you're probably not using the server (you can configure this, but it defaults to early morning -- ~ 3AM I believe).  Note, however, that this means writes to the array are NOT fault-tolerant until after the mover has run and moved the data to the protected array.

 

Personally, I prefer to know that all data I write to my UnRAID array is fault-tolerant after it's written; so I do not use a cache drive.    But if you want the higher speed writes, that's how you can get them.