Hi,

currently i'm not pleased with the performance i get from my unraid server when writing to the SSD Cache drive.

 

Unraid: 6.2

SSD: Samsung 840 Pro, 256GB

Disk Usage: 34% (1VM)

Format: btrfs

Network: 1Gbit/s

 

So a normal file Transfer with SMB share currently looks like:

 

It starts with maximum speed (limited by network) and drops off to 20-50mb/s after ~600 to 1000MB of data.

 

Model family:	Samsung based SSDs
Device model:	Samsung SSD 840 PRO Series
Serial number:	S1ATNEAD560077H
LU WWN device id:	5 002538 55033e322
Firmware version:	DXM05B0Q
User capacity:	256,060,514,304 bytes [256 GB]
Sector size:	512 bytes logical/physical
Rotation rate:	Solid State Device
Device:	In smartctl database [for details use: -P show]
ATA version:	ACS-2, ATA8-ACS T13/1699-D revision 4c
SATA version:	SATA 3.1, 6.0 Gb/s (current: 6.0 Gb/s)
Local time:	Fri Oct 28 11:34:55 2016 CEST
SMART support:	Available - device has SMART capability.
SMART support:	Enabled
SMART overall-health:	Passed

 

Smart seems to be fine.

5	Reallocated sector count	0x0033	100	100	010	Pre-fail	Always	Never	0
9	Power on hours	0x0032	098	098	000	Old age	Always	Never	5542 (7m, 16d, 22h)
12	Power cycle count	0x0032	097	097	000	Old age	Always	Never	2302
177	Wear leveling count	0x0013	096	096	000	Pre-fail	Always	Never	141
179	Used rsvd block count tot	0x0013	100	100	010	Pre-fail	Always	Never	0
181	Program fail count total	0x0032	100	100	010	Old age	Always	Never	0
182	Erase fail count total	0x0032	100	100	010	Old age	Always	Never	0
183	Runtime bad block	0x0013	100	100	010	Pre-fail	Always	Never	0
187	Uncorrectable error count	0x0032	100	100	000	Old age	Always	Never	0
190	Airflow temperature cel	0x0032	067	055	000	Old age	Always	Never	33
195	ECC error rate	0x001a	200	200	000	Old age	Always	Never	0
199	CRC error count	0x003e	100	100	000	Old age	Always	Never	0
235	POR recovery count	0x0012	099	099	000	Old age	Always	Never	224
241	Total lbas written	0x0032	099	099	000	Old age	Always	Never	25772364270

 

And some Disk-Log Information from last Reboot.

Oct 27 20:01:13 UNRAID kernel: ata6: SATA max UDMA/133 abar m2048@0xdf12c000 port 0xdf12c380 irq 125
Oct 27 20:01:13 UNRAID kernel: ata6: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
Oct 27 20:01:13 UNRAID kernel: ata6.00: READ LOG DMA EXT failed, trying unqueued
Oct 27 20:01:13 UNRAID kernel: ata6.00: failed to get NCQ Send/Recv Log Emask 0x1
Oct 27 20:01:13 UNRAID kernel: ata6.00: ATA-9: Samsung SSD 840 PRO Series, S1ATNEAD560077H, DXM05B0Q, max UDMA/133
Oct 27 20:01:13 UNRAID kernel: ata6.00: 500118192 sectors, multi 16: LBA48 NCQ (depth 31/32), AA
Oct 27 20:01:13 UNRAID kernel: ata6.00: failed to get NCQ Send/Recv Log Emask 0x1
Oct 27 20:01:13 UNRAID kernel: ata6.00: configured for UDMA/133
Oct 27 20:01:13 UNRAID kernel: ata6.00: Enabling discard_zeroes_data
Oct 27 20:01:13 UNRAID kernel: sd 6:0:0:0: [sdg] 500118192 512-byte logical blocks: (256 GB/238 GiB)
Oct 27 20:01:13 UNRAID kernel: sd 6:0:0:0: [sdg] Write Protect is off
Oct 27 20:01:13 UNRAID kernel: sd 6:0:0:0: [sdg] Mode Sense: 00 3a 00 00
Oct 27 20:01:13 UNRAID kernel: sd 6:0:0:0: [sdg] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Oct 27 20:01:13 UNRAID kernel: ata6.00: Enabling discard_zeroes_data
Oct 27 20:01:13 UNRAID kernel: sdg: sdg1
Oct 27 20:01:13 UNRAID kernel: ata6.00: Enabling discard_zeroes_data
Oct 27 20:01:13 UNRAID kernel: sd 6:0:0:0: [sdg] Attached SCSI disk
Oct 27 20:01:13 UNRAID kernel: BTRFS: device fsid a1ce48b5-34ee-4745-8a40-67ca67d04a29 devid 1 transid 32809 /dev/sdg1
Oct 27 20:01:20 UNRAID emhttp: Samsung_SSD_840_PRO_Series_S1ATNEAD560077H (sdg) 250059064
Oct 27 20:01:20 UNRAID emhttp: import 30 cache device: sdg
Oct 27 20:01:22 UNRAID kernel: BTRFS info (device sdg1): disk space caching is enabled

 

I'm not sure if the device itself can't handle the speed.

Any help to analyse the problem is appreciated.

 

Regards,

Trynn

Are you regularly trimming your SSD? If not install the Dynamix trim plugin and schedule it to run daily.

You nailed it. Thanks

Did TRIM manually first with:

 

fstrim -v /mnt/cache

 

Transferspeed is now steady @max Network.

Just installed dynamix trim pluggin as adviced. (https://lime-technology.com/forum/index.php?topic=36543.0)

 

But not sure how to schedule it to run "daily". There are no options i can find in the web-GUI :-)

Settings -> Scheduler -> SSD TRIM Settings

<3 got it