Having upgraded from TrueNAS Core 12.0-U1 to 12.0-U2, the system starts kernel panicking when there's HDD load of any kind (Scrubbing, copying files, etc.)
System Details:
IBM X3650 M3,
IBM M1015 HBA (LSI 2008 chipset) connected to IBM SAS expander.
Debug logs attached.
Note: Reverting to 12.0-U1 fixed the issue and the system is stable again.
Problem/Justification
None
Impact
None
SmartDraw Connector
Katalon Manual Tests (BETA)
Activity
Show:
William Gryzbowski
May 7, 2021 at 5:36 PM
Hi Robert,
Unfortunately there doesn't seem to be much we can do here ourselves other than you manually applying that tunable.
We will verify if there is somewhere in the documentation we can note it.
Robert Johnston
April 16, 2021 at 6:25 AM
Having added the loader tunable and upgraded to U3, this all appears to be working just fine. Odd that the tunable caused problems like this, but I'm glad it's resolved.
Alexander Motin
April 5, 2021 at 7:57 PM
(edited)
The only change related to old mps(4) LSI HBAs in U2 was reduction of maximum I/O size from 1MB to 128KB by adding hw.mps.max_io_pages=32 loader tunable. It was done after several people reported firmware hangs I guessed triggered by use of larger I/Os in 12. I have no clue so far how setting the tunable can cause that, but you ma try to return some back by setting hw.mps.max_io_pages loader tunable via the WebUI to -1, which is the default value.
William Gryzbowski
March 22, 2021 at 12:35 PM
Sorry it may be a while, we have limited resources and people are working on all different kind of issues with its own priority.
Unfortunately Jira tickets from community have to get lower priority sometimes.
Robert Johnston
March 19, 2021 at 8:08 PM
It's been over a week with no movement on this. Did the new logs not help?
Having upgraded from TrueNAS Core 12.0-U1 to 12.0-U2, the system starts kernel panicking when there's HDD load of any kind (Scrubbing, copying files, etc.)
System Details:
IBM X3650 M3,
IBM M1015 HBA (LSI 2008 chipset) connected to IBM SAS expander.
Debug logs attached.
Note: Reverting to 12.0-U1 fixed the issue and the system is stable again.