failing to boot pass mdadm monitor

Peter M. Petrakis peter.petrakis at
Mon Mar 12 21:11:14 UTC 2012

On 03/12/2012 04:13 PM, Asif Iqbal wrote:
> I am failing to boot this server x4270 pass the mdadm --monitor.
> Installed lucid amd64. This is a first install.
> details:
> It boots all the way in recovery mode. what gives?

I would say your problems began once your partition detection went

[   26.603469] sd 6:0:3:0: [sdd] 585937500 512-byte logical blocks: (300 GB/279 GiB)
[   26.603599] GPT:Primary header thinks Alt. header is not at the end of the disk.
[   26.603600] GPT:585937498 != 585937499
[   26.603602] GPT:Alternate GPT header not at the end of the disk.
[   26.603603] GPT:585937498 != 585937499

which is coming from fs/partitions/efi.c

  494        if (le64_to_cpu(agpt->my_lba) != lastlba) {
  495                printk(KERN_WARNING
  496                       "GPT:Alternate GPT header not at the end of the disk.\n");
  497                printk(KERN_WARNING "GPT:%lld != %lld\n",
  498                        (unsigned long long)le64_to_cpu(agpt->my_lba),
  499                        (unsigned long long)lastlba);
  500                error_found++;
  501        }

The math for lastlba seems correct, 585937500 - 1ULL, a quick google shows
the raw size is consistent with what you have. So the question is how
did 585937498 get computed? Would someone with some more experience with
GPT partitions care to comment?

It doesn't look like your system successfully recovered from find_valid_gpt
or we would have seen "Alternate GPT is invalid, using primary GPT." or
"Primary GPT is invalid, using alternate GPT." in the logs. Those partitions,
or what's left of them, is being presented to mdadm for assembly.

This part is really weird.

[   27.929744] mdadm: sending ioctl 1261 to a partition!
[   27.935401] mdadm: sending ioctl 1261 to a partition!
[   27.941905] mdadm: sending ioctl 1261 to a partition!

Where 1261 is #define __NR_set_mempolicy              1261

I don't think that has any business being sent to a partition. At the moment,
I can't explain how this event, and the GPT fault could be related.

[   28.679619] md1: detected capacity change from 0 to 146360172544
[   28.684287]  md1: unknown partition table
[   28.709647] mdadm: sending ioctl 1261 to a partition!
[   28.713255] mdadm: sending ioctl 1261 to a partition!
Begin: Running /scripts/local-premount ...
[   29.082318] EXT3-fs: INFO: recovery required on readonly filesystem.
[   29.088849] EXT3-fs: write access will be enabled during recovery.
[   29.252295] kjournald starting.  Commit interval 5 seconds
[   29.252387] EXT3-fs: recovery complete.
[   29.265115] EXT3-fs: mounted filesystem with ordered data mode.

I really don't know what shape your backing store is in, noting your
second post, I'm surprised your disks are readable.

So... what changed? Have you upgraded system firmware recently,
kernel upgrades, or anything at all really? I see you have an external
storage enclosure, has that seen any changes either?


More information about the ubuntu-server mailing list