"Could not prepare Boot variable: No such file or directory" when cobbler imaging server
by str123 from LinuxQuestions.org on (#6QGY9)
Hi,
I just joined the community to see if anyone had any ideas on this issue.
We've been successfully imaging Dell R350 servers using cobbler in a production environment. The target OS is CentOS7.9. We're now trying to image the new Dell R360 using the same cobbler and same CentOS7.9. The Dell servers are very similar, with the only things changing being items that Dell updated since the last generation of server. The R360 imaging process fails during bootloader.write as seen below, however, there is a prompt given to continue regardless of the error. If you continue, then the imaging process does succeed, and you can successfully boot to CentOS7.9. The only minor issue I see post imaging process is a stale boot entry "Unavailable: CentOS Linux" in the UEFI boot options. The BIOS mode is set to UEFI (the only option).
The anaconda.log file shows:
Code:21:33:32,028 INFO anaconda: Installing boot loader
21:33:32,028 INFO anaconda: boot loader stage1 target device is sda1
21:33:32,028 INFO anaconda: boot loader stage2 target device is sda2
21:33:32,029 DEBUG anaconda: new default image: <pyanaconda.bootloader.LinuxBootLoaderImage object at 0x7f624e06d310>
21:33:32,134 INFO anaconda: bootloader.py: used boot args: crashkernel=auto rhgb quiet
21:33:33,492 ERR anaconda: bootloader.write failed: failed to set new efi boot target. This is most likely a kernel bug.
The program.log file shows:
Code:21:33:32,112 INFO program: Running... efibootmgr
21:33:32,124 INFO program: BootCurrent: 0000
21:33:32,124 INFO program: Timeout: 0 seconds
21:33:32,124 INFO program: BootOrder: 0002,0000,0004,0001,0003
21:33:32,124 INFO program: Boot0000* Embedded NIC 1 Port 1 Partition 1
21:33:32,124 INFO program: Boot0001* DVD-ROM DP61N
21:33:32,125 INFO program: Boot0002* CentOS Linux
21:33:32,125 INFO program: Boot0003* EFI Fixed Disk Boot Device 1
21:33:32,125 INFO program: Boot0004* CentOS Linux
21:33:32,125 DEBUG program: Return code: 0
21:33:32,125 INFO program: Running... efibootmgr -c -w -L CentOS -d /dev/sda -p 1 -l \EFI\centos\shimx64.efi
21:33:32,134 INFO program: Could not prepare Boot variable: No such file or directory
21:33:32,134 DEBUG program: Return code: 5Interestingly, the R350 Boot000x entries show as "Boot000x CentOS", whereas the R360 Boot000x entries show as "Boot000x CentOS Linux".
When I get into the anaconda shell and run "efibootmgr -c -w -L CentOS -d /dev/sda -p 1 -l \EFI\centos\shimx64.efi" manually, it returns the same error as above. The \EFI\centos\shimx64.efi file exists as does /dev/sda. Not sure what the "No such file or directory" is referring to.
The servers have HW RAID controllers with two SSDs in RAID1. I tried imaging with just one drive in non-RAID and got the same error.
I'm now wondering if the OS is too old and some incompatibility is finally rearing its head...
I just joined the community to see if anyone had any ideas on this issue.
We've been successfully imaging Dell R350 servers using cobbler in a production environment. The target OS is CentOS7.9. We're now trying to image the new Dell R360 using the same cobbler and same CentOS7.9. The Dell servers are very similar, with the only things changing being items that Dell updated since the last generation of server. The R360 imaging process fails during bootloader.write as seen below, however, there is a prompt given to continue regardless of the error. If you continue, then the imaging process does succeed, and you can successfully boot to CentOS7.9. The only minor issue I see post imaging process is a stale boot entry "Unavailable: CentOS Linux" in the UEFI boot options. The BIOS mode is set to UEFI (the only option).
The anaconda.log file shows:
Code:21:33:32,028 INFO anaconda: Installing boot loader
21:33:32,028 INFO anaconda: boot loader stage1 target device is sda1
21:33:32,028 INFO anaconda: boot loader stage2 target device is sda2
21:33:32,029 DEBUG anaconda: new default image: <pyanaconda.bootloader.LinuxBootLoaderImage object at 0x7f624e06d310>
21:33:32,134 INFO anaconda: bootloader.py: used boot args: crashkernel=auto rhgb quiet
21:33:33,492 ERR anaconda: bootloader.write failed: failed to set new efi boot target. This is most likely a kernel bug.
The program.log file shows:
Code:21:33:32,112 INFO program: Running... efibootmgr
21:33:32,124 INFO program: BootCurrent: 0000
21:33:32,124 INFO program: Timeout: 0 seconds
21:33:32,124 INFO program: BootOrder: 0002,0000,0004,0001,0003
21:33:32,124 INFO program: Boot0000* Embedded NIC 1 Port 1 Partition 1
21:33:32,124 INFO program: Boot0001* DVD-ROM DP61N
21:33:32,125 INFO program: Boot0002* CentOS Linux
21:33:32,125 INFO program: Boot0003* EFI Fixed Disk Boot Device 1
21:33:32,125 INFO program: Boot0004* CentOS Linux
21:33:32,125 DEBUG program: Return code: 0
21:33:32,125 INFO program: Running... efibootmgr -c -w -L CentOS -d /dev/sda -p 1 -l \EFI\centos\shimx64.efi
21:33:32,134 INFO program: Could not prepare Boot variable: No such file or directory
21:33:32,134 DEBUG program: Return code: 5Interestingly, the R350 Boot000x entries show as "Boot000x CentOS", whereas the R360 Boot000x entries show as "Boot000x CentOS Linux".
When I get into the anaconda shell and run "efibootmgr -c -w -L CentOS -d /dev/sda -p 1 -l \EFI\centos\shimx64.efi" manually, it returns the same error as above. The \EFI\centos\shimx64.efi file exists as does /dev/sda. Not sure what the "No such file or directory" is referring to.
The servers have HW RAID controllers with two SSDs in RAID1. I tried imaging with just one drive in non-RAID and got the same error.
I'm now wondering if the OS is too old and some incompatibility is finally rearing its head...