0 Replies Latest reply on Nov 25, 2014 10:06 AM by waveform

    RHEL 7.0, stuck on fsck!

    waveform Novice


       I am trying to build a RHEL-7.0 based VM on VMware Studi 2.6 but am hitting the following error.  Does anyone have any tips to fix/workaround this error?


      /dev/sda1: The filesystem size (According to the superblock) is 131040 blocks
      The physical size of the device is 128488 blocks
      Either the superblock or the partition table is likely to be corrupt!
      /dev/sda1: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
      (i.e., without -a or -p options)




      The error appears when the provisioned VM PXE-boots prior to accessing its kickstart file.  The RHEL 7.0 template is based on /opt/vmware/etc/build/templates/redhat/6/0_x86_64 with minimal adjustments (e.g. ISO location).


      To try to workaround the error, I added fsck.mode=skip to the boot parameters, but this doesn't get me much farther.  Shortly further in the pre-PXE boot, a network-related error occurs:


      RTNETLINK answers: File exists
      Warning: Could not boot.




      I have tried a couple of things to troubleshoot:


      On one hand, I have tried running through the following steps manuallychanging vmware-vdiskmanager's target type to -t 0 so that I can mount and test a single file-system fileand the resulting fs looks good.

       


      Ran system shell command "unzip /opt/vmware/share/build/vmdk/grubvmdk.zip 2>/tmp/YgJ0d40I3c"
      system shell command returned: exit 0, signal 0, core 0
      system shell command stdout:
      Archive: /opt/vmware/share/build/vmdk/grubvmdk.zip
      inflating: grub.vmdk
      inflating: grub-flat.vmdk

      Ran system shell command "losetup -f 2>/tmp/qr59D0F46C"
      system shell command returned: exit 0, signal 0, core 0
      Ran system shell command "losetup -o 32256 /dev/loop7 /opt/vmware/www/build/XXX/installer/grub-flat.vmdk 2>/tmp/dYKkDP6tmf"
      system shell command returned: exit 0, signal 0, core 0
      Ran system shell command "mount -t auto /dev/loop7 /opt/vmware/www/build/XXX/installer/mount 2>/tmp/49kPcLnSdH"
      system shell command returned: exit 0, signal 0, core 0
      Copying file /mnt/V46135-01.iso/images/pxeboot/vmlinuz to /opt/vmware/www/build/XXX/installer/mount/vmlinuz:
      Copying file /mnt/V46135-01.iso/images/pxeboot/initrd.img to /opt/vmware/www/build/XXX/installer/mount/initrd.img
      Ran system shell command "umount -l /opt/vmware/www/build/XXX/installer/mount 2>/tmp/jX5HFB12x4"
      system shell command returned: exit 0, signal 0, core 0
      Ran system shell command "losetup -d /dev/loop7 2>/tmp/xPmtOoUXIG"

      ...

      Creating virtual disk "YYY"
      Ran system shell command "/usr/bin/vmware-vdiskmanager -r "/opt/vmware/www/build/XXX/installer/grub.vmdk" -t 1 "/opt/vmware/www/build/XXX/va/YYY-disk1.vmdk" 2>/tmp/l3jYvkrCBh"
      system shell command returned: exit 0, signal 0, core 0
      Ran system shell command "/usr/bin/vmware-vdiskmanager -x 8192Mb "/opt/vmware/www/build/XXX/va/YYY-disk1.vmdk" 2>/tmp/zSQhcuLbIq"




      OTOH, I have intercepted the grub.vmdk-based boot disk of a (different, RHEL-6.3-based) working build, and the working disk also fails an fsck:

       

      # fsck /dev/sdb1

      fsck from util-linux-ng 2.17.2

      e2fsck 1.41.12 (17-May-2010)

      The filesystem size (according to the superblock) is 131040 blocks

      The physical size of the device is 128488 blocks

      Either the superblock or the partition table is likely to be corrupt!

      Abort<y>? yes


      So, on one hand, a recreation of the failing disk looks good when I examine it and, OTOH, a succeeding disk looks bad if I examine it pre-PXE-boot!  I am stumped!  SOS!