Opened 5 weeks ago

#833 new defect

Issues when restoring RHEL7 /boot under XFS format. Just getting GRUB RESCUE prompt

Reported by: jdaraya Owned by: bruno
Priority: high Milestone: 3.3.0
Component: mondo Version: 3.2.2
Severity: blocker Keywords: grub rescue mondo xfs boot
Cc:

Description

Hello guys,

I have several years using mondo and I would like to thank you for such an amazing application! you have help me a lot! I had been trough several situations and have learn a lot from you using mondorescue packages.

In our company we develop at least 2 or 3 times a year Disaster Recovery Exercises, and we have just updated one of our environments to RHEL7.

On the last DRE we couldn't restore two images. We had all the same versions (latest) than other servers running RHEL6 and RHEL5 but this time we just got a "GRUB RESCUE>" prompt.

Testing the images on a local server I got the same error and I tried even the expert mode and partitioning manually everything and restoring the whole data, but it continued showing the same error.

Before continuing I will post some useful information for you:


OS efi afio Mondo perl-ProjectBuilder? perl-Mondo Mindi Mindi BusyBox? Mindi debuginfo Size Mondo line Result


RHEL 6.8 yes afio-2.5-1.rhel6.x86_64 mondo-3.2.2-1.rhel6.x86_64 perl-ProjectBuilder?-0.12.7-1.rhel6.noarch perl-MondoRescue?-3.2.2-1.rhel6.noarch mindi-3.0.2-1.rhel6.x86_64 mindi-busybox-1.21.1-1.rhel6.x86_64 mindi-debuginfo-3.0.1-1.rhel6.x86_64 2.3G mondoarchive -Oi -9 -N -I / -E '/app' -K 99 -f /dev/sda -S /tmp/mondo -T /tmp/mondo -d /Linux_Unix_Images/redhat/wlv-dff-mslap1 -s 4480m -p mslap1-rescuecd Worked fine


RHEL 6.8 yes afio-2.5-1.rhel6.x86_64 mondo-3.2.2-1.rhel6.x86_64 perl-ProjectBuilder?-0.12.7-1.rhel6.noarch perl-MondoRescue?-3.2.2-1.rhel6.noarch mindi-3.0.2-1.rhel6.x86_64 mindi-busybox-1.21.1-1.rhel6.x86_64 mindi-debuginfo-3.0.1-1.rhel6.x86_64 2.3G mondoarchive -Oi -9 -N -I / -E '/app' -K 99 -f /dev/sda -S /tmp/mondo -T /tmp/mondo -d /Linux_Unix_Images/redhat/wlv-dff-mslap2 -s 4480m -p mslap2-rescuecd Worked fine after EFI modification


RHEL 6.8 yes afio-2.5-1.rhel6.x86_64 mondo-3.2.2-1.rhel6.x86_64 perl-ProjectBuilder?-0.12.7-1.rhel6.noarch perl-MondoRescue?-3.2.2-1.rhel6.noarch mindi-3.0.2-1.rhel6.x86_64 mindi-busybox-1.21.1-1.rhel6.x86_64 mindi-debuginfo-3.0.1-1.rhel6.x86_64 2.2G mondoarchive -Oi -9 -N -l GRUB -f /dev/sda -I / -S /tmp/mondo -T /tmp/mondo -d /Linux_Unix_Images/redhat/wlv-dff-msldb4 -s 4480m -p msldb4-rescuecd Worked fine


RHEL 6.8 no afio-2.5-1.rhel6.x86_64 mondo-3.2.2-1.rhel6.x86_64 perl-ProjectBuilder?-0.12.7-1.rhel6.noarch perl-MondoRescue?-3.2.2-1.rhel6.noarch mindi-3.0.2-1.rhel6.x86_64 mindi-busybox-1.21.1-1.rhel6.x86_64 mindi-debuginfo-3.0.1-1.rhel6.x86_64 1.8G mondoarchive -Oi -9 -N -I / -S /tmp/mondo -T /tmp/mondo -d /Linux_Unix_Images/redhat/wlv-dff-msldb5 -s 4480m -p msldb5-rescuecd Worked fine but no graphic console


RHEL 7.3 no afio-2.5-1.rhel7.x86_64 mondo-3.2.2-1.rhel7.x86_64 perl-ProjectBuilder?-0.14.4-1.rhel7.noarch perl-MondoRescue?-3.2.2-1.rhel7.noarch mindi-3.0.2-1.rhel7.x86_64 mindi-busybox-1.21.1-1.rhel7.x86_64 mindi-debuginfo-3.0.1-1.rhel7.x86_64 1.9G mondoarchive -Oi -9 -N -I / -E '/app' -K 99 -f /dev/sda -S /tmp/mondo -T /tmp/mondo -d /Linux_Unix_Images/redhat/wlv-siapp1-prd -s 4480m -p siapp1-rescuecd Issues


RHEL 7.3 no afio-2.5-1.rhel7.x86_64 mondo-3.2.2-1.rhel7.x86_64 perl-ProjectBuilder?-0.14.4-1.rhel7.noarch perl-MondoRescue?-3.2.2-1.rhel7.noarch mindi-3.0.2-1.rhel7.x86_64 mindi-busybox-1.21.1-1.rhel7.x86_64 mindi-debuginfo-3.0.1-1.rhel7.x86_64 2.5G mondoarchive -Oi -9 -N -I / -E '/app' -K 99 -f /dev/sda -S /tmp/mondo -T /tmp/mondo -d /Linux_Unix_Images/redhat/wlv-siapp2-prd -s 4480m -p siapp2-rescuecd Issues


As I said I was getting the same error on a LAB developed locally, then I tried several things including setting Grub parameters in order to have the right partition booted. None of those worked. Since troubleshooting mondorescue is something I have done a couple times before I tried a work around and after the "successful" restoration I booted again from the DVD and saw that all the file systems had the information on them, even boot. Then I realized that the information was been set on place but by some reason the app is not seeing the boot partition. Tried running the mkinitrd and fdisk the /dev/sda in order to check everything, but always got to the same place.

Something that looked extrange for me was that both partitions /dev/sda1 (boot) and /dev/sda2 (LVM) were marked as bootable partitions then I though something was going on when seting the bootable partition, Then I decided to run manually a mkfs.xfs getting to the same place. What make notice that, as far as I know, thats a RHEL7 characteristic(xfs boot partition), then I reformatted the partition as ext3 and booted the image as INTERACTIVE, allowing to only recover /boot data on the already partitioned and formated partition (voiding those steps on the interactive mode) and "voilà" IT WORKED!

Then, the major issue here seems to be that mondo is not recognizing or setting the format correctly(xfs) for boot, (I don't know where to fix that)

Hope this looooong story helps and you can have it easily fixed. I bet for sure it will be really easy for you guys to do it.

Thanks a lot for the opportunity and keep the great work! I really appreciate it.

Best regards,

  1. Diego Araya R.

Attachments (1)

mondo issue summary.txt (2.6 KB) - added by jdaraya 5 weeks ago.
Details of OS version and packages installed with the line executed and result when restoring.

Download all attachments as: .zip

Change History (1)

Changed 5 weeks ago by jdaraya

Details of OS version and packages installed with the line executed and result when restoring.

Note: See TracTickets for help on using tickets.