Repeated file system corruption

I am suffering from repeated filesystem corruption issues when using the linux turney installations (I have tested both the base package and the Jenkins packages.).

The corruption manifests as a remount of the root file system read only. From this point, the filesystem can sometimes be recovered using fsck after rebooting into maintenance mode, but often maintenance mode can't boot and I get the busybox prompt or the fsck utility itself is corrupt as it cannot run - from this point attaching to another linux system and running fsck usually fixes the problem, but the corruption always re-occurs again.

I have used the vmdk images directly as well as installing from the iso into ext4/lvm, ext4 and ext3 filesystems. I used VirtualBox running on at least 2 different PCs as well as running the images in a vmware cloud - all will at some point become corrupt. Note all the installations were run at least once or set up in virtualbox.

The logs tend to mention issues such as "jenkins kernel: [] e2label[]: segfault at... ip .. sp erro" or "attempt to access beyond end of device, aborting journal on device sda1"

Anyone any ideas?

Jeremy Davis's picture

Although not sure if that would be your issue seeing as you have tried on alternate machine.

I have heard one other report of a similar situation and IIRC it was traced back to corruption of the underlaying FS (IIRC it was hardware issues). So that's probably where I'd start looking. I have been running many TKL apliances in VirtualBox and VMware server (from both ISO and VM builds) on Win XP, Server 2003, Win 7 and various forms of Ubuntu and have only ever had one case of VM FS corruption (which was solved by fsck from a live ISO).

Also make sure that you are shutting it down properly. IIRC you need to install acpi-support (ie 'apt-get install acpi-support') to get it all working nicely...

Post new comment