bootimage problems with IPMI client
Verfasst: 21 Jun 2017, 23:16
I have a board here I couldn't install with opsi for a long while. It always resets in the windows setup while "expanding files". Until somebody asked if I had looked in the ipmi log. Which I hadn't.
There were lots of entries like this:
Somehow the ipmi watchdog got activated. The ipmi watchdog is seperate from the motherboard watchdog, and I think it is mandatory since ipmi 1.0.
part from the linux logfile on the machine:
The ipmi watchdog is set by the bootimage somehow. From a different host I queried the bmc controller and asked about the status for the watchdog:
After factory reset:
When the linux bootimage started this was the status:
When the linux bootimage stops, it sets the watchdog:
After the reboot the counter continues running. The countdown always ran out during the "expanding files" part of the windows setup. This was a coincidence. After running the bootimage, I entered the bios setup, waited for the reset, and afterwards the setup finally worked.
The windows setup never continued after the reset by the ipmi watchdog, because the opsi batch file refers to the wrong drive.
Tried with opsi-linux-bootimage.noarch 20170508-1.1 and a couple before that.
So uhh, turning off the ipmi watchdog in the bootimage seems the most logical step to do?
There were lots of entries like this:
Code: Alles auswählen
152,2017/05/09 15:15:30, #0x00,OS Boot,C: Boot Completed - Assertion
153,2017/05/09 15:16:42, #0xca,Watchdog 2,Timer Interrupt - Assertion
154,2017/05/09 15:16:43, #0xca,Watchdog 2,Hard Reset - Assertion
part from the linux logfile on the machine:
Code: Alles auswählen
[7] [May 17 19:21:45] >>> [ 10.180028] intel_idle: lapic_timer_reliable_states 0xffffffff (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.180031] ipmi message handler version 39.2 (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.180045] ipmi device interface (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.180062] ipmi_si IPI0001:00: ipmi_si: probing via ACPI (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.180086] ipmi_si IPI0001:00: [io 0x0ca4] regsize 1 spacing 1 irq 0 (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.180597] ipmi_si: Adding ACPI-specified kcs state machine (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.181103] IPMI System Interface driver. (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.181606] ipmi_si: probing via SMBIOS (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.182105] ipmi_si: SMBIOS: io 0xca2 regsize 1 spacing 1 irq 0 (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.182612] ipmi_si: Adding SMBIOS-specified kcs state machine (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.183120] ipmi_si: probing via SPMI (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.183620] ipmi_si: SPMI: io 0xca2 regsize 1 spacing 1 irq 0 (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.184130] ipmi_si: SPMI-specified kcs state machine: duplicate (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.184820] ipmi_si: Trying ACPI-specified kcs state machine at i/o address 0xca4, slave address 0x0, irq 0 (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.458143] ipmi_si IPI0001:00: The BMC does not support clearing the recv irq bit, compensating, but the BMC needs to be fixed. (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 10.586475] ipmi_si IPI0001:00: Error clearing flags: cc (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 11.490233] ipmi_si IPI0001:00: Found new BMC (man_id: 0x002a7c, prod_id: 0x0886, dev_id: 0x20) (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 11.490828] ipmi_si IPI0001:00: IPMI kcs interface initialized (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 11.493810] IPMI Watchdog: driver initialized (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 11.494355] Copyright (C) 2004 MontaVista Software - IPMI Powerdown via sys_reboot. (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 11.562154] IPMI poweroff: ATCA Detect mfg 0x2A7C prod 0x886 (Posix.py|845)
[7] [May 17 19:21:45] >>> [ 11.562712] IPMI poweroff: Found a chassis style poweroff function (Posix.py|845)
After factory reset:
Code: Alles auswählen
Watchdog Timer Use: Reserved (0x00)
Watchdog Timer Is: Stopped
Watchdog Timer Actions: No action (0x00)
Pre-timeout interval: 0 seconds
Timer Expiration Flags: 0x00
Initial Countdown: 0 sec
Present Countdown: 0 sec
Code: Alles auswählen
Watchdog Timer Use: SMS/OS (0x04)
Watchdog Timer Is: Stopped
Watchdog Timer Actions: No action (0x00)
Pre-timeout interval: 0 seconds
Timer Expiration Flags: 0x00
Initial Countdown: 10 sec
Present Countdown: 10 sec
Code: Alles auswählen
Watchdog Timer Use: SMS/OS (0x44)
Watchdog Timer Is: Started/Running
Watchdog Timer Actions: Hard Reset (0x01)
Pre-timeout interval: 0 seconds
Timer Expiration Flags: 0x00
Initial Countdown: 120 sec
Present Countdown: 111 sec
The windows setup never continued after the reset by the ipmi watchdog, because the opsi batch file refers to the wrong drive.
Tried with opsi-linux-bootimage.noarch 20170508-1.1 and a couple before that.
So uhh, turning off the ipmi watchdog in the bootimage seems the most logical step to do?