VMware Cloud Community
srodenburg
Expert
Expert

SSD Wear-leveling indicator / S.M.A.R.T. Data

Hello,

I Read The Friendly Manual but I can't seem to find readouts for S.M.A.R.T. data and the SSD Wear-leveleling Indicator. I use LSI 9207-8i HBA's so ESXi has direct control over the disks.

Kind regards,

Steve

Reply
0 Kudos
5 Replies
crosdorff
Enthusiast
Enthusiast

Have you tried:

# esxcli storage core device list

# esxcli storage core device smart get -d device

Where device is a value found in step 1.


VMware KB: ESXi S.M.A.R.T. health monitoring for hard drives

Reply
0 Kudos
srodenburg
Expert
Expert

Thanks. Yes I have tried that. But the output is not very helpfull:

[root@esx02:~] esxcli storage core device smart get -d naa.50011731002b819c

Parameter                     Value  Threshold  Worst

----------------------------  -----  ---------  -----

Health Status                 OK     N/A        N/A 

Media Wearout Indicator       N/A    N/A        N/A 

Write Error Count             0      N/A        N/A 

Read Error Count              0      N/A        N/A 

Power-on Hours                N/A    N/A        N/A 

Power Cycle Count             515    N/A        N/A 

Reallocated Sector Count      N/A    N/A        N/A 

Raw Read Error Rate           N/A    N/A        N/A 

Drive Temperature             34     N/A        N/A 

Driver Rated Max Temperature  N/A    N/A        N/A 

Write Sectors TOT Count       N/A    N/A        N/A 

Read Sectors TOT Count        N/A    N/A        N/A 

Initial Bad Block Count       N/A    N/A        N/A 

The "Wearout Indicator" does not work as you can see. Most counters don't work. The nodes are all fully HCL compliant:  LSI 9702-8i  with FW 19 and driver v19.00.00.00.1vmw  and the SSD is a Sandisk 200GB that is also on the HCL.

That why i asked. Maybe i'm doing something wrong.

Does the "Wearout Indicator" work for anyone else? At all?

If i stick that SSD in a Windows server which has the SanDisk SAS SSD Toolkit installed, i see everything. So the drive does have all the counters (as one would expect from an Enterprise SAS SSD).

Reply
0 Kudos
zdickinson
Expert
Expert

Is there a SANDisk utility?  There is realssd for Micron and that show's the wear level/expected life.  Thank you, Zach.

Reply
0 Kudos
pvlasek
Contributor
Contributor

It works little bit better for us with Intel S3700, S3500, S3610 SSD connected to LSI 9300 adapter:

Parameter                 Value  Threshold  Worst

----------------------------  -----  ---------  -----

Health Status             OK N/A    N/A
Media Wearout Indicator   99 0      99
Write Error Count         N/AN/A    N/A
Read Error Count          N/AN/A    N/A
Power-on Hours            1000      100
Power Cycle Count         1000      100
Reallocated Sector Count  1000      100
Raw Read Error Rate       N/AN/A    N/A
Drive Temperature         1000      100
Driver Rated Max Temperature 85 0      71
Write Sectors TOT Count   1000      100
Read Sectors TOT Count    N/AN/A    N/A
Initial Bad Block Count   10090     100
Reply
0 Kudos
srodenburg
Expert
Expert

There is the SSD Toolkit for Enterprise SSD drives for Windows. It does show everything. On the SanDisk Support-site, i found this page under the drive's download-area.

There is a separate CLI tool for Linux:  Optimus Ascend SSD Drive Firmware Webpage (includes tools for Windows and Linux)

The accompanying documentation "Enterprise Storage Manager CLI User Guide", mentiones "using the scli tool under VMware" but the link to the chapter is broken and there is not a word anywhere about VMware. Very strange document...


The Linux download has a "ubuntu" folder and a "generic" folder. I tried all the variants, both 32bit and 64bit and they all Segfault when run. The help does show, but as soon as you do a "scli show all" command, all variants simply crash with a seg-fault.


The "generic 64 bit version" does not crash when using it in the following way:

If i specify the disk in a way that "i think" the tool expects according to the documentation (read: linux like), it does not segfault then, but says that the device is not supported:


/dev/disks/naa.1234567890 : Device type not supported.

/vmfs/devices/disks/naa.1234567890 : Device type not supported.

vmhba1:C0:T0:L0 : Device type not supported.

etc. etc. etc.  nothing works so far.

So that's it. I have found no usable software from Sandisk for VMware ESXi. The CLI tool's documentation suggests that it should work but so far it doesn't. I have no clue how the tool expects the disk-device path to be written in the syntax. The documentation is abysmally poor.


But honestly, I expected that all of this stuff would "just work" when selecting a HBA, a SSD, drivers and firmwares that are on the HCL.

Reply
0 Kudos