SOLVED Hard disk temperatures not plotted on Reporting section

pixel8383

Explorer
Joined
Apr 19, 2018
Messages
80
I am using FreeNAS from years and today I noticed that the hard disk temperatures are not plotted on the graphs of the Reporting section. The last values I can see on the graphs are from June 2019 (I don't remember which FreeNAS version was released at the time), after that the graphs are blank. I am using a script scheduled to run every week which returns the right temperatures so the information is still available to the system. The version I am running now is TrueNAS-12.0-U2.1 but the system has been updated over time on every stable release.

How can I fix it?
 

pixel8383

Explorer
Joined
Apr 19, 2018
Messages
80
To everyone looking for an answer, I found it. In order for disk temperature monitoring to work, S.M.A.R.T. must be enabled for disk, and also either "HDD Standby" should be set to "Always On", or "Force HDD Standby" feature should be enabled.

Thanks anyway to the community. :smile:
 

theaddies

Contributor
Joined
Mar 28, 2015
Messages
105
I am running 12.0-U8.1 and I have "HDD Standby" set to "Always On" for all disks in my pool. I don't see a temperature plot. I must be missing something. Please help.
 

dspeed

Cadet
Joined
Jun 8, 2022
Messages
7
Please remove the [SOLVED] flag; the general problem does not appear to be resolved.

I am running TRUENAS 13.0-U2 on a Supermicro X10SL7-F motherboard; the drives are connected to the sas controller interface. No temps shown in Reporting tab for 2 installed 6TB WD Red Plus drives. All of the disk parameters are shown as null.

Drives are set as always on and SMART enabled under the Storage/Disks tab.

This feature in TRUENAS 13.0-U2 appears to be working on another X10SL7-F the drives are connected to the sas controller interface.

I will try this with some other drives after I'm done with the initial burn-in on these two drives (On reflection, I'm thinking that I have observed this before...).

Maybe a BIOS or IPMI setting?
 

Samuel Tai

Never underestimate your own stupidity
Moderator
Joined
Apr 24, 2020
Messages
5,398
@dspeed, please don't revive necro threads, but start your own. Note SMART is a SATA technology, which doesn't apply to SAS drives. Some SAS controllers can pass through SMART info from attached SATA drives, but not all. What controller are you using? If it's a RAID controller in JBOD mode, that's even worse.
 

dspeed

Cadet
Joined
Jun 8, 2022
Messages
7
@Samuel Tai, thanks for your prompt attention to this matter.

Please note that I was the second user this month posting about this issue. This forum is invaluable to the TRUENAS community. My comment was to add value to the original March 2021 post in that the suggested solution does not solve the issue in all cases.

Will happily start a new thread after I get a chance look at a few items. Searching and reading a few other (mostly older) posts did not turn up an obvious answer. For clarification, my system(s) are set to use English and I did not look at any elements (other than cpu and disk) in the Reporting tab in my Test system.

But my comment was made to correct an error, in that the suggested settings from 03/2021 DO NOT fix the problem in all cases.


I understand that SATA is not SAS. My other X10SL7-F system is using SATA drives on the SAS controller and the temps are being reported. Again, I did not look at any other values on my Production system in the Reporting tab (was not in 'testing mode').

Supermicro specifies the X10SL7-F controller as a "SAS2 (6Gbps) via LSI 2308". A link to some info on the 2308 here:

https://www.servethehome.com/lsi-sas-2308-raid-controller-hba-information-listing/

While I did peek at the controller settings once, I made no changes. I do not think mine is currently configured as a RAID, but (now that I see it is possible) I will look more closely (after current burn-in activities complete). I have had my test system filled with a number of older drives and was able to make pools and vdeves at will with no issues.

FWIW, I have an HP204 SAS controller and cables on order to try my hand at 12 Gbps SAS for my next system.

At present I've reached 9.8 Gbps over DAC through my Mikrotik CRS305 transferring from 6 5400 rpm SATA drives on 6 Gbps SAS (?) interfaces. Is this indicative of a problem? My 10G quest has been a challenge with an odd mix of 1-10+ year old gear (mostly used). But it is fun to me.

Dave

Production system (today)
TRUENAS 13.0-U2 (updated from 12, 11, ...)
X10SL7-F
Xeon E3-1230-v3
32GB 1600 ECC RAM
250 gb M2 cache
x520-1 10GBE
2x 6 TB WD reds (cmr)
6x 3 TB WD reds (cmr)

Test system (today)
TRUENAS 13.0-U2 (fresh install)
X10SL7-F
Xeon E3-1241-v4
32GB 1600 ECC RAM
500 gb M2 cache
x520-1 10GBE
2x 6 TB SATA WD red plus (cmr)
1x 12 TB SAS WP Arsenal [on order for testing a de-dupe setup]

FreeNAS user since 9. (something) iirc ... its been awhile; my old login got lost long ago.
 

Samuel Tai

Never underestimate your own stupidity
Moderator
Joined
Apr 24, 2020
Messages
5,398
The 2308 can run in both IR (RAID) and IT mode. Only IT mode is supported with TrueNAS.
 

theaddies

Contributor
Joined
Mar 28, 2015
Messages
105
I have a Supermicro X11SPM-F motherboard with 64 GB ECC ram. I am running 7 x 10 TB drives in raid Z2. I upgraded to Version: TrueNAS-13.0-U2 and nothing changed. I get no reports at all.
 

dspeed

Cadet
Joined
Jun 8, 2022
Messages
7
Again, thanks for the feedback.

TBH, I generally let upgrades slide awhile unless there is an identifiable problem, because it seems that every time I upgrade something breaks.

There may be an issue with my controller setup, but I don't see where it is affecting the temp (and other) readings in the Reports tab. I reach this conclusion because the problem noted above still exists with the 2 drives being tested connected as ada1 and ada2 (as a mirror) to the SATA only sockets on the [Test] X10SL7-F and nothing on the LSI controller. (The boot SSD was always connected to ada0). I have not tested relative performance (it's all 10G at the moment and it was hybrid 1G/10G before.

And, BTW, the cpu/core temps show up on the Dashboard tab, changing in relation to the system load, so I suspect they are measured values.

Indeed, the LSI chip identifies itself as SMC2308-IR; it has a firmware of 15:00:00:00-IR. Both motherboards have the same LSI configuration. Any pointers to changing the firmware (I think I've seen it for cards, but never looked for built-in) or if there is indeed a need to change? Mine has been chugging along happily for the past 6+ years in its current configuration, but I'm always up for better.

Just shutdown the Production machine and looked at the LSI setup; It appears the same as Test after a reasonable scrutiny; will compare side to side laBTW, what's up with this forum software? I can't seem to cut and paste to fix the flow.lease pardon the flow; ter, but I'm pretty sure they are identical. All of my 8 storage pool drives (in 1 2 disk mirror and 1 6 disk ZFS2 pool) are connected via the LSI controller.

The older [Production] system was purchased new by me some time ago (8-10 yrs) and I do not recall changing the LSI settings. The Test system was purchased used last month.

I'll open a new thread on this in a day or two. Will include bits of this (edited) as backgound.



Dave
 

dspeed

Cadet
Joined
Jun 8, 2022
Messages
7
I found a wealth of info here on updating the x10sl7-f, Will collect links and post current notes. After I get it to work....
 

douglasg14b

Dabbler
Joined
Nov 26, 2017
Messages
26
@Samuel Tai "necroing threads" is a non-logical stance, as fragmented information is more difficult to find than information gathered in a small number of threads. What is the purpose of not necroing threads? Is it just because there is some stigma, or is there reasoning behind it, and what is the mitigation for the knock-on effects of encouraging information fragmentation?

This is worth calling out, as search engines will index these threads at the top, and all you get are dead threads and dead ends becuase of this "no necroing" lemming-like stance.



That said, I also have this problem, temperature doesn't show up for me there are no graphs at all.

In my case it seems that SMART fails to start, does your SMART service start successfully @dspeed
 

dspeed

Cadet
Joined
Jun 8, 2022
Messages
7
@Samuel Tai "necroing threads" is a non-logical stance, as fragmented information is more difficult to find than information gathered in a small number of threads. What is the purpose of not necroing threads? Is it just because there is some stigma, or is there reasoning behind it, and what is the mitigation for the knock-on effects of encouraging information fragmentation?

This is worth calling out, as search engines will index these threads at the top, and all you get are dead threads and dead ends becuase of this "no necroing" lemming-like stance.



That said, I also have this problem, temperature doesn't show up for me there are no graphs at all.

In my case it seems that SMART fails to start, does your SMART service start successfully @dspeed

At present, I don't run SMART automatically (mine is not a 24/7 system).

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

Beg pardon, but I got lost down the rabbit hole investigating the "Reporting Tab not displaying drive temps on the X10SL7-F mobo". (Used to do real research and hate to post w/o peer review <grin>.)

In this process, I built another NAS to enable testing both with and without the flashed onboard controller. I see no difference in reporting. BTW, my initial results show no obvious performance changes (tho I will admit that all variables were not held constant...).

I believe I am seeing a difference in the reporting of pools created under different versions of Free/True NAS; I have one pool that I believe was created under V9, but I do not have that era in my logs.

This week I'm looking for a few good (and cheap) 12gbps HBAs, and appreciating the wealth of information scattered throughout this forum.
While I've been a casual user for enough years for my original login to have gotten 'lost', I've really learned a lot in the past couple of months by RTFM. Whatta concept.

Kudos, folks.
 
Top