LSI 9206-16e Temperature Threshold flags

[wallacearizona]

Hello I am seeking some help on a new setup using the LSI 9206-16e.  I was able to get the card flashed to IT mode with the BIOS but for some reason my MPT2SAS keeps filling up with temp flag warnings.

 

Apr 15 09:20:04 kernel: mpt2sas_cm0: Temperature Threshold flags 0 exceeded for Sensor: 0 !!!

Apr 15 09:20:04 kernel: mpt2sas_cm0: Current Temp In Celsius: 80

Apr 15 09:20:24 kernel: mpt2sas_cm1: Temperature Threshold flags 0 exceeded for Sensor: 0 !!!

Apr 15 09:20:24 kernel: mpt2sas_cm1: Current Temp In Celsius: 60

 

I have read the temperature and air flow requirements for the card and I am within specs for the card.

 

Can someone tell me how to change the Temperature Threshold Flags or how to silence the messages in the logs?

 

Attached is my Diagnostics.

 

Thanks Everyone! 

wallaceserver-diagnostics-20210415-1143.zip

[Mrobert]

Hi,

 

any luck with that?

I have the same flood of messages in my log from the beginning and it's slowly starting to be a pain...

[DeathStar Darth]

Having just installed my first LSI 9201-16e and seen this in my logs, I went digging.

 

It looks like this is coming directly from the driver

/**

* _scsih_temp_threshold_events - display temperature threshold exceeded events

* @ioc: per adapter object

* @event_data: the temp threshold event data

* Context: interrupt time.

*

* Return nothing.

*/

static void

_scsih_temp_threshold_events(struct MPT2SAS_ADAPTER *ioc,

Mpi2EventDataTemperature_t *event_data)

{

if (ioc->temp_sensors_count >= event_data->SensorNum) {

printk(MPT2SAS_ERR_FMT "Temperature Threshold flags %s%s%s%s"

"exceeded for Sensor: %d !!!\n", ioc->name,

((event_data->Status & 0x1) == 1) ? "0 " : " ",

((event_data->Status & 0x2) == 2) ? "1 " : " ",

((event_data->Status & 0x4) == 4) ? "2 " : " ",

((event_data->Status & 0x8) == 8) ? "3 " : " ",

event_data->SensorNum);

printk(MPT2SAS_ERR_FMT "Current Temp In Celsius: %d\n",

ioc->name, event_data->CurrentTemperature);

}

}

 

I don't know about you, but I don't want to be compiling my own driver

I will be ordering a Noctura (or two) to somehow strap to what appeared to be a really rather good heatsink - putting the case sides on sees the reported temps rapidly rise...suspect I am going to need to add another extractor to the case as well to cope with the extra heat the Noctura will (hopefully) suck away from the LSI.

[wallacearizona]

20 minutes ago, DeathStar Darth said:

Having just installed my first LSI 9201-16e and seen this in my logs, I went digging.

 

It looks like this is coming directly from the driver

/**

* _scsih_temp_threshold_events - display temperature threshold exceeded events

* @ioc: per adapter object

* @event_data: the temp threshold event data

* Context: interrupt time.

*

* Return nothing.

*/

static void

_scsih_temp_threshold_events(struct MPT2SAS_ADAPTER *ioc,

Mpi2EventDataTemperature_t *event_data)

{

if (ioc->temp_sensors_count >= event_data->SensorNum) {

printk(MPT2SAS_ERR_FMT "Temperature Threshold flags %s%s%s%s"

"exceeded for Sensor: %d !!!\n", ioc->name,

((event_data->Status & 0x1) == 1) ? "0 " : " ",

((event_data->Status & 0x2) == 2) ? "1 " : " ",

((event_data->Status & 0x4) == 4) ? "2 " : " ",

((event_data->Status & 0x8) == 8) ? "3 " : " ",

event_data->SensorNum);

printk(MPT2SAS_ERR_FMT "Current Temp In Celsius: %d\n",

ioc->name, event_data->CurrentTemperature);

}

}

 

I don't know about you, but I don't want to be compiling my own driver

I will be ordering a Noctura (or two) to somehow strap to what appeared to be a really rather good heatsink - putting the case sides on sees the reported temps rapidly rise...suspect I am going to need to add another extractor to the case as well to cope with the extra heat the Noctura will (hopefully) suck away from the LSI.

Yes, the only post I have seen about these temp messages is to remove the bios from the card but on my machine it will not initialize with Unraid without the bios present.  
 

Yes this card runs extremely hot.  I created a custom vent shroud with 2 60mm fans removing heat away and still runs in the 60’s .

 

If anyone with the log problem has a bit of coding experience could we parse the logs of these messages?

[DeathStar Darth]

On 10/21/2022 at 8:51 PM, wallacearizona said:

If anyone with the log problem has a bit of coding experience could we parse the logs of these messages?

 

I am VERY inexperienced with the sed command but putting something like 

sed -i '/\(mpt2sas_cm[12]: Temperature Threshold flags\|mpt2sas_cm[12]: Current Temp In Celsius\)/d' /var/log/syslog

in a user script to run every hour (or frequency of your choosing) will remove the entries from the syslog to make it easier to read (ie without the noise from the entries that you can't control/wish to ignore), but still keep any other mpt2sas entries.

I've posted this on https://github.com/techie2000/unraid/tree/main/boot/config/plugins/user.scripts/scripts/Delete mpt2sas noise from syslog for future reference

[petepaccione]

I ran into this issue also, for those who use a remote rsyslog server for logs, if you want to avoid using a cron-sed script, edit your /etc/rsyslog.d/UNRAID.log and add this:

 

$template unraid, "/var/log/unraid.log"

 

# Filter out specific noisy messages before logging
if ($fromhost-ip == '192.168.UNRAID-IP-HERE' and ($msg contains 'mpt2sas_cm1:') or ($msg contains 'mpt2sas_cm0:')) then stop

# If the message is from the specific host and has not been stopped by the previous rule, log it
if ($fromhost-ip == '192.168.UNRAID-IP-HERE') then -?unraid
& stop

 

You can add all kinds of "remove if found" filters by using the ($msg contains 'whatever lines you want to remove') - you can use booleans between the parenthesis