[6.9.0-RC1 and 6.9.0-RC2] - Cache disk temp will not save

Also 6.9.0-RC1 here.
I was able to successfully update cache and cache 2  for those settings, I had previously set them in earlier releases, but this release had reset those settings.

2 hours ago, tjb_altf4 said:

Also 6.9.0-RC1 here.
I was able to successfully update cache and cache 2  for those settings, I had previously set them in earlier releases, but this release had reset those settings.

Weird, all mine had also reset but i cannot change those settngs on only the cache disk, and on further investigation i cannot change the utilisation thresholds either.

Here's the bug in action

 

https://i.gyazo.com/61888ec40c67157ae4a53e025dfa30a7.mp4

33 minutes ago, Bolagnaise said:

Here's the bug in action

Dark blue numbers on a dark grey background is a combination that just doesn't work for my eyes. I see what you're saying though. Maybe your diagnostics would reveal more.

32 minutes ago, John_M said:

Dark blue numbers on a dark grey background is a combination that just doesn't work for my eyes. I see what you're saying though. Maybe your diagnostics would reveal more.

Log shows this when attempting to set temp or utilisation limits.

Dec 14 16:58:52 Tower kernel: mdcmd (92): set md_num_stripes 1280
Dec 14 16:58:52 Tower kernel: mdcmd (93): set md_queue_limit 80
Dec 14 16:58:52 Tower kernel: mdcmd (94): set md_sync_limit 5
Dec 14 16:58:52 Tower kernel: mdcmd (95): set md_write_method

 

Diagnostics attached

tower-diagnostics-20201214-1658.zip

Edited December 14, 2020 by Bolagnaise

I too am experiencing issues with being unable to specify the temperature threshold for any of the disks in my array and pools, after updating to 6.9.0-rc1.

 

This becomes a bit more than annoying when NVMe SSDs are involved due to the average temps for those being higher than the default threshold values.

 

So far, it seems that the only workaround is to change the default values at the OS level.

 

Maybe this thread could be escalated from a status of "annoyance" to "minor"?

Thoughts?

6 hours ago, Mathew R. said:

I too am experiencing issues with being unable to specify the temperature threshold for any of the disks in my array and pools, after updating to 6.9.0-rc1.

 

This becomes a bit more than annoying when NVMe SSDs are involved due to the average temps for those being higher than the default threshold values.

 

So far, it seems that the only workaround is to change the default values at the OS level.

 

Maybe this thread could be escalated from a status of "annoyance" to "minor"?

Thoughts?

Agreed, ill update the thread. My NVME drives are also causing issues with warnings.

Updated to RC2, problem still exists, cannot save temp or drive utilisation limits for the Cache drive.

I just updated to 6.9.0-rc2 and drive temperatures are no longer being displayed. Instead, I get an asterisks symbol where the drive temperature would normally be displayed.

 

As well, I too am still unable to edit the drive temperature values as seen and reported in 6.9.0-rc1.

As to your first problem, this is a known issue with RC2. Manually spin up your drives (even if they're already spun up, from what I understand) by clicking the little status LED symbol on the left. The temperatures should display correctly after this.

Thanks for the tip -Deadalus

Just tested on my system, I can update SATA SSDs with no issues in cache and test pool HDDs, but I dont have any NVME devices in my system.

 

Looks like values are stored here in boot/config.

 

root@Tower:/boot/config# cat smart-one.cfg 
[disk1]
smEvents="5|187|197|198"
[SSD_PLUS_480GB_]
warning="80"
hotTemp="50"
maxTemp="60"
[CT500MX500SSD1_]
hotTemp="55"
maxTemp="65"
[ST96812AS_]
hotTemp="55"
maxTemp="65"

 

Disk name as serial number after underscore.

Edited December 21, 2020 by SimonF

Interesting...,

In my "smart-one.cfg", I can see the correct temperature values for the cache pools I have, but there are no disks specified for each of the temperature value sets (e.g., where it would display the disk model and serial number, it only displays the pool assignment, "cache, cache2, scratch, scratch2).

 

 

In my "smart-one.cfg", I can see my configuration for the unassigned disks I use with ZFS

[Samsung_SSD_860_EVO_250GB_xxx]
hotTemp="55"
maxTemp="65"
[Samsung_SSD_860_EVO_250GB_xxx]
hotTemp="55"
maxTemp="65"
[Samsung_SSD_860_EVO_500GB_xxx]
hotTemp="55"
maxTemp="65"
[Samsung_SSD_860_EVO_250GB_xxx]
hotTemp="55"
maxTemp="65"
[Samsung_SSD_970_EVO_Plus_1TB_xxx]
hotTemp="55"
maxTemp="65"

but the unraid system doesn't care and complains all the time using the default settings.

I just set UD Dev to 20 and 30.

 

Dashboard is correct.

 

 

But I didn't get notification until I changed the defaults so agree its not using device specific ones.

 

@limetech UDs are not using device specific temps definitions for notifications only the defaults. But dashboard is using specific values.

Edited December 21, 2020 by SimonF

yep, that is correct. The dashboard is fine... the notification is the trouble

Ok, just following up, your right, my smart-one.cfg shows the correct temps but the dashboard does not update and still throws temp and usage alarms.

 

[SPCC_M_2_PCIe_SSD_A44907990E8900022704]
hotTemp="55"
maxTemp="60"
[ST4000VN008-2DR166_ZGY2T6RW]
hotTemp="50"
maxTemp="60"
[ST4000VN008-2DR166_ZDH3DQ5V]
hotTemp="50"
maxTemp="60"
[WDC_WD40EFRX-68N32N0_WD-WCC7K1KCA1LR]
hotTemp="50"
maxTemp="60"
[WDC_WD30EZRX-00DC0B0_WD-WMC1T0247433]
hotTemp="50"
maxTemp="60"
[Samsung_SSD_840_Series_S19HNSAD555889P]
hotTemp="50"
maxTemp="60"
[SanDisk_SDSSDA120G_160547400593]
hotTemp="50"
maxTemp="60"
 

 

 

 

 

Just updated to the 6.9.0 Stable build and I am still unable to get the temp settings to stick.

Is anyone else also still experiencing this issue?

27 minutes ago, Mathew R. said:

Just updated to the 6.9.0 Stable build and I am still unable to get the temp settings to stick.

Is anyone else also still experiencing this issue?

I'm currently dumbfounded...

The issue seems to now be specific to FireFox (although I had never thought previously to test with a different browser prior to the 6.9.0 Stable release).

e.g., When updating the temp values in FireFox, they appear to save initially, but then revert randomly.

In Chrome, when updating the values, almost immediately, you can see a visual difference. The default values are denoted in a blue toned color rather than being a grayish color in FireFox.

 

I stumbled upon this workaround when looking for reports in the stable build threads and someone else had reported that using a different browser worked.

 

Potentially the problem could also be related to cached data/cookies. I don't really know. Just glad that I am not having the issue anymore.

 

Good luck to you all.

13 hours ago, Mathew R. said:

The issue seems to now be specific to FireFox

 

Agreed. I'm seeing the same with Firefox 85.0.2 (macOS). It also applies twhen setting custom SMART parameters, as well as temperature thresholds. Chrome 88.0.4324.192 works fine.

Ok having similar issues but i think i discovered my answer with the above thread... please correct me if i'm wrong but changing the individual drive limits ONLY affects the dashboard color values etc... and DOES NOT change the notifications / emails about a drive being outside of limits.  IE if i want my unraid to stop emailing me when my NVME gets a little hot, i have to adjust the GLOBAL variables?

 

I'm hoping thats the case cuz it would explain why i keep getting notified that my nvme is running a little hot even after adjusting the individual drive limits... but i'm also hoping i'm doing something wrong cuz i would love to not get the emails about the NVME based on it's individual settings and not have to change the global variable just for one drive that runs a touch hotter.

 

On 6/11/2021 at 2:41 PM, DirkRockface said:

Ok having similar issues but i think i discovered my answer with the above thread... please correct me if i'm wrong but changing the individual drive limits ONLY affects the dashboard color values etc... and DOES NOT change the notifications / emails about a drive being outside of limits.  IE if i want my unraid to stop emailing me when my NVME gets a little hot, i have to adjust the GLOBAL variables?

 

I'm hoping thats the case cuz it would explain why i keep getting notified that my nvme is running a little hot even after adjusting the individual drive limits... but i'm also hoping i'm doing something wrong cuz i would love to not get the emails about the NVME based on it's individual settings and not have to change the global variable just for one drive that runs a touch hotter.

 

This was most definitely not the case in prior releases. I have always set my cache drives to a higher temp to avoid the false notifications. I was unable to do that in 6.9, even on 6.9.2. However, it did seem to be a browser issue for me as well. I just tried it in Edge, was previously using Chrome, and it appears to have updated and stuck in the GUI. I will report back if I still get a temp alarm email notification below what I set next time my drives are working hard.

Still having this issue in 6.9.2 as well. Tried multiple browsers, no change.

Same here, 6.9.2 and NVME temperature setting won't save - normal disks will though.

 

Tried firefox and edge

+1

Also SMART values are "empty"...some are displayed, when Controller-type "Automatic" or "NVMe" gets selected.

However, this setting also does not stick.

 

unRAID 6.9.2, NVMe: Crucial Force MP600 (1TB), cache pool of 2 devices