Closed ghost closed 1 month ago
For some more context, here's what OL9.4 looks like when this procedure is done on a VM and the installer succeeds in starting (apologies for the screenshot format):
Did you try to do an in-place upgrade from CentOS 7 to Oracle Linux 8 ? Here the steps: https://blogs.oracle.com/scoter/post/upgrade-centos-7-to-oracle-linux-8 Once done, you can then upgrade to Oracle Linux 9 with Leapp: https://docs.oracle.com/en/operating-systems/oracle-linux/9/leapp/
Hi @scoter-oracle,
We don't really have the resources to change the procedure now and we are changing too much in the stack to ensure operation after doing the upgrade via leapp. There are dozens of these machines outside our geographical location making the upgrade procedure a bit more complicated as they have no OOB interface, and we would really like to keep supporting them.
Since Anaconda supports loading the stage2 off of a disk, this procedure works on newer hardware/VM's, and AlmaLinux 9.4 is able to operate as expected on this older hardware, I believe there is an issue with the OL9.4 vmlinuz/initrd.img combo and I'm happy to provide more information to resolve it. I can get my hands on the hardware again within a week, and I'll add rd.debug
.
Thanks
Oracle Linux customers, please file your issue at https://support.oracle.com
Thanks for filing an issue with Oracle Linux.
GitHub Issues is not an official support channel and we don't offer product support here. If you're not yet an Oracle Linux customer, consider signing up at https://linux.oracle.com.
Even if you're not a customer, if we can confirm that an issue is a bug we will do our best to fix it and to update this issue once it has been fixed. We don't guarantee a fix or feedback and for now, we will close this issue. If you have Oracle Linux support, please use support.oracle.com to report issues.
Hi,
Sure. Then just so you're aware there is an issue specific to the OL9 Anaconda installer that does not/cannot mount the filesystem when specifying inst.stage2=hd:LABEL=$PARTITION_LABEL
(substituting $PARTITION_LABEL for installer
in my case) on certain hardware.
Using the same kernel parameters with AlmaLinux 9.4 and Rocky Linux 9.4 both succeed. OL9.3 and OL9.4 are unable to mount and timeout/drop to the dracut shell.
I don't expect a response, just trying to be helpful.
Here is the hardware info:
description: Desktop Computer
product: APU2 (2 GB)
vendor: PC Engines
version: 1.0
serial: 1408231
width: 64 bits
capabilities: smbios-2.7 dmi-2.7 smp vsyscall32
configuration: boot=normal chassis=desktop sku=2 GB
*-core
description: Motherboard
product: APU2
vendor: PC Engines
physical id: 0
version: 1.0
serial: 1408231
*-firmware
description: BIOS
vendor: coreboot
physical id: 0
version: 4.0.7
date: 02/28/2017
size: 1MiB
capacity: 8MiB
capabilities: pci pcmcia upgrade bootselect acpi
*-cpu:0 DISABLED
description: CPU [empty]
vendor: AuthenticAMD
physical id: 4
version: AMD GX-412TC SOC
configuration: cores=4
*-memory
description: System memory
physical id: 1
size: 1920MiB
*-cpu:1
product: AMD GX-412TC SOC
vendor: Advanced Micro Devices [AMD]
physical id: 2
bus info: cpu@0
version: 22.48.1
size: 1GHz
capacity: 1GHz
width: 64 bits
capabilities: fpu fpu_exception wp vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp x86-64 constant_tsc rep_good acc_power nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb bpext ptsc perfctr_llc cpb hw_pstate ssbd vmmcall bmi1 xsaveopt arat npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold overflow_recov cpufreq
configuration: microcode=117637381
*-pci:0
description: Host bridge
product: Family 16h (Models 30h-3fh) Processor Root Complex
vendor: Advanced Micro Devices, Inc. [AMD]
physical id: 100
bus info: pci@0000:00:00.0
version: 00
width: 32 bits
clock: 33MHz
[...]
*-sata
description: SATA controller
product: FCH SATA Controller [IDE mode]
vendor: Advanced Micro Devices, Inc. [AMD]
physical id: 11
bus info: pci@0000:00:11.0
logical name: scsi1
version: 40
width: 32 bits
clock: 66MHz
capabilities: sata pm ahci_1.0 bus_master cap_list emulated
configuration: driver=ahci latency=64
resources: irq:19 ioport:4010(size=8) ioport:4020(size=4) ioport:4018(size=8) ioport:4024(size=4) ioport:4000(size=16) memory:feb25000-feb253ff
*-disk
description: ATA Disk
product: Samsung SSD 860
physical id: 0.0.0
bus info: scsi@1:0.0.0
logical name: /dev/sda
version: 4B6Q
serial: S4CJNF0N202657J
size: 232GiB (250GB)
capabilities: partitioned partitioned:dos
configuration: ansiversion=5 logicalsectorsize=512 sectorsize=512 signature=0002e532
*-volume:0
description: Linux filesystem partition
physical id: 1
bus info: scsi@1:0.0.0,1
logical name: /dev/sda1
logical name: /boot
capacity: 1GiB
capabilities: primary bootable
configuration: mount.fstype=xfs mount.options=rw,relatime,attr2,inode64,logbufs=8,logbsize=32k,noquota state=mounted
*-volume:1
description: Linux filesystem partition
physical id: 2
bus info: scsi@1:0.0.0,2
logical name: /dev/sda2
logical name: /installer
capacity: 2GiB
capabilities: primary
configuration: mount.fstype=xfs mount.options=rw,relatime,attr2,inode64,logbufs=8,logbsize=32k,noquota state=mounted
*-volume:2
description: Linux LVM Physical Volume partition
physical id: 3
bus info: scsi@1:0.0.0,3
logical name: /dev/sda3
serial: DBH3UV-8aul-zViP-NsDU-0OwT-Lrgx-CiP9ea
size: 229GiB
capacity: 229GiB
capabilities: primary multi lvm2
00:11.0 SATA controller: Advanced Micro Devices, Inc. [AMD] FCH SATA Controller [IDE mode] (rev 40) (prog-if 01 [AHCI 1.0])
Subsystem: Advanced Micro Devices, Inc. [AMD] FCH SATA Controller [IDE mode]
Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 19
I/O ports at 4010 [size=8]
I/O ports at 4020 [size=4]
I/O ports at 4018 [size=8]
I/O ports at 4024 [size=4]
I/O ports at 4000 [size=16]
Memory at feb25000 (32-bit, non-prefetchable) [size=1K]
Capabilities: [60] Power Management version 3
Capabilities: [70] SATA HBA v1.0
Kernel driver in use: ahci
Kernel modules: ahci
Thanks and goodbye.
Hello,
We are upgrading many physical machines from CentOS 7 to OL9.4. They have less than 3GB of RAM, so loading the stage2 Anaconda install.img resources over the network is not possible. We do not have physical access to these machines, so we're trying to come up with a way to upgrade them remotely without doing an exclusively network based upgrade procedure.
We came up with a procedure to workaround this memory limitation by loading all installer resources (pxeboot/vmlinuz, pxeboot/initrd.img, install.img, and the kickstart) off the disk via a specific partition label (/dev/sda2 LABEL=installer in the below output).
Unfortunately, these machines seem to run into a bug that is not present with other EL distributions. We tested this same procedure with the same kernel parameters with AlmaLinux 9.4 and everything worked as expected. This procedure also works with other (newer) hardware and VM's on Oracle Linux 9.4, so I believe that this is a hardware specific issue.
The installer is started via kexec. It is unable to load the install.img file to start Anaconda and times out. We never see the kernel messages that the filesystem gets mounted. However, I am able to scan the disk for partitions and mount it in the dracut/rescue installer environment that I get dropped into by hand.
I haven't attempted this with RHEL 9.4 or Rocky, but the fact that this works with AlmaLinux 9.4 leads me to believe there might be a bug in the initrd.img for the OL9.4 installer (for specific/older hardware).
Here's the output of an attempt:
I'll reiterate that this exact procedure with the same kernel parameters works with AlmaLinux 9.4 on the problematic hardware, and OL9.4 with this exact procedure/kernel parameters on more modern hardware/VM's works as well.
Thanks