search

SUSE-Enceladus / azure-li-services

Azure Large Instance Services
GNU General Public License v3.0
7 stars 0 forks source link

SLES12-SP4 need manual intervention to boot. #232

Closed prtyag closed 4 years ago

prtyag commented 4 years ago

Hi, Current VLI test image SLES12-SP4-SAP-Azure-VLI-BYOS.x86_64-0.0.23-Devel-Build33.2.raw is not booting up on a VLI system. I can see message on the console saying "reboot: system halted". Console output screenshot is below:

error_screen_1 error_screen_2

After this point I reboot the system and it came up.

schaefi commented 4 years ago

I'll give it a test. So far I see the xfs_repair failed with a segmentation fault which is the reason why you got stuck

I'll test the appliance on a VM and see if I can reproduce this

schaefi commented 4 years ago

The boot in a VM worked flawlessly. Could this maybe be a problem of your storage system and/or the LUN you bumped the image onto. See the XFS rootfs is not clean which it is if you only look at the image file. Also the xfs_repair failed with a segmentation fault. That really points to a problem with the storage or the drivers (fb) or both.

schaefi commented 4 years ago

Very strange is also that it boots on a second attempt, if the xfs root is broken I would have expected the same situation on reboot

prtyag commented 4 years ago

The boot in a VM worked flawlessly. Could this maybe be a problem of your storage system and/or the LUN you bumped the image onto. See the XFS rootfs is not clean which it is if you only look at the image file. Also the xfs_repair failed with a segmentation fault. That really points to a problem with the storage or the drivers (fb) or both.

If the drivers(fb) or storage has a problem then we would have got the same response from other images as well but that's not the case other images are booting up fine.

prtyag commented 4 years ago

Very strange is also that it boots on a second attempt, if the xfs root is broken I would have expected the same situation on reboot

Yes I'm also wondering the same.

schaefi commented 4 years ago

Hmm, there is really little I can do here. Can you repeat that xfs_repair segmentation fault any time ? Meaning is this a safe reproducer ? If yes I suggest you open a bugzilla entry against the xfsprogs maintainer (ailiopoulos@suse.com) and include the segfault error as shown in your output. The image can serve as a reproducer on your end.

schaefi commented 4 years ago

@prtyag I'm closing this on behalf of the bug report you created against xfs. As there is little we can do here I think this history can be closed. Thanks