Can't read CTR while initalizing i8042+Failed to start OpenSSH server d.-> unable to login

[tamariz]

EQUIPMENT: HP PROLIANT DL380 GEN10 PLUS

OPERATING SYSTEM: RHEL 8.9

-0-

AFTER ELECTRICAL CUT THE SERVER DOES NOT BOOT SUCCESSFULLY: AS SOON AS THE SYSTEM PROMPT IS SHOWN, THE KEYBOARD SEEMS TO BE UNRESPONSIVE - NO INPUT PROVIDED-. 

WE RESTART FROM ILO (v.5) AND THE SEQUENCE WAS THE FOLLOWING ONE:

01. We get the following messages:

• [0.070382] x86/cpu: SGX disabled by BIOS.
• [2.171788] i8042: Can't read CTR while initalizing i8042

02. The system then continues the boot process normally (it seems):

   

03:  Subsequently we are asked to introduce the LUKS2 decryption key and later on we get the following error message: “Failed to start OpenSSH server daemon”.

04. And finally, we get the system prompt in ILO remote console

However, it seems the keyboard is unresponsive, no input can be provided and therefore we can not login into the system.

According to:

https://support.hpe.com/hpesc/public/docDisplay?docLocale=en_US&docId=emr_na-a00041662en_us

Advisory: Linux - Kernel Error Message "i8042: Can't Read CTR while Initializing i8042" May Be Displayed When Booting HPE Moonshot/Edgeline, HPE ProLiant Gen9, or HPE ProLiant Gen10 Servers

The CTR issue does not seem to be a problem, however it seems there are several users that already had this problem. (?)

-0-

On the other hand, OpenSSH failure does not seem to be a problem preventing us to login into the system (?).

-0-

As part of the troubleshooting:

• As part of the troubleshooting, we entered the RHEL emergency mode by pressing the ‘e’ key while the system is booting.
• This allows to edit the Grub ‘linux’ line, replacing the “ro” with “rw init=/sysroot/bin/sh” and pressing CTRL + X to boot the system.
• After boot, ‘:/# ‘ prompt is shown, and we are able to use systemctl as systemd is enabled and running. However, we cannot see the partitions of the system yet, and the services shown are not the ones present on the system when normal boot is completed succesfully. To mount the root system, we executed the command:

:/# chroot /sysroot

• This allowed me to access the root partition. However, after this command, I cannot use ‘systemctl’ commands as systemd is no recognized on the system. As far as we have read on fórums, systemd and systemctl are only available if systemd is PID 1.
• How can we mount the rest of partitions apart from the /root? As /var and /var/log/ partitions are separated and not mounted, I cannot access them to check the logs.
• Is it posible to mount the partitions in emergency mode and still be able to use systemd / systemctl?
• When in emergency mode, journalctl is not showing any boot information, as commands like ‘journalctl -b 1’ ‘journalctl -b 3’ are empty. How can we access this information from emergency mode? Is there a possibility to access this information on normal boot before the system crashes? (apparently, it crashes when login prompt is shown)

-0-

What we can do? What we are doing in a wrong way? Any help will be really appreciated.

Many thanks in advance.

Carlos.

[Sham82]

Hello @tamariz,

Thank you for your POST.

https://support.hpe.com/hpesc/public/docDisplay?docId=a00041662en_us&docLocale=en_US
as per the document :
"The "Can't read CTR while initializing i8042" message can be safely ignored. This occurs because the Linux kernel is attempting to locate a PS/2 controller that is not attached."

You can also check - https://access.redhat.com/discussions/d7d7ee0d-2549-454a-b577-62251d707ca5

Regards
HPE

[tamariz]

Good aftetrnoon, thanks for your answer, but:

the redhat link you refer to has been generated by me (we open the case to RHEL as well in case)

On the other hand, in ILO we have:

Server Critical Fault (Service Information: Input Powe Loss, Power Supply, (Power Supply 1 (03h) Power Supply 2 (03h))

Event Class: 0x14

Event Code: 0xA

http://www.hpe.com/support/class0x0014code0x000a-gen10

There were two electrical cuts in the facilities and that's the reason of this message. Considering a visual inspection we did not see any hardware issue but we have not been able to run a fcsk, for example, up to now. On the other hand, our system is securized at BIOS level, hard disks level and OS level (in order to be qualify our system). I'm wondering if makes sense to carry out some tests removing securization partially.

Regards,

Carlos.

[Sham82]

Hello @tamariz,

For further investigation from H/W end , we would suggest to log a support case with HPE
with latest AHS logs.

Regards
HPE