On 10/15/2014 11:53 AM, Don Armstrong wrote:
On Wed, 15 Oct 2014, Jape Person wrote:
From dmesg
...[ 4.853751] systemd-gpt-auto-generator[154]: Failed to determine
partition table type of /dev/sda: Input/output error
[ 4.854298] systemd[151]:
/lib/systemd/system-generators/systemd-gpt-auto-generator failed with error
code 1.
... and later on ...
[12650.204616] systemd-gpt-auto-generator[7555]: Failed to determine
partition table type of /dev/sda: Input/output error
...
I looked in the BTS and couldn't even find a package named
systemd-gpt-auto-generator, much less a bug that had been filed for
it. I guess it's a routine or function name?
It's part of systemd; it generates rules to mount partitions from GPT
partition tables without needing to express them in /etc/fstab. [See man
systemd-gpt-auto-generator for details.]
Now that's weird, or maybe it's just me. I tried to look for manpages
for systemd-gpt-auto-generator, and I'd swear I was told "No manual
entry for..."
Sven Joachim told me to check the manpages, and I looked at man systemd,
which gave me an online reference for "Generators Specifications" which
wasn't helpful at all.
But I'm now seeing documentation when I type "man
systemd-gpt-auto-generator", so I'm guessing I made a typo earlier on
and didn't even notice in the output from the man request.
The drive came originally from Lenovo (T520i) with and MSDOS parititon
table. I just used the standard partition scheme provided by the
netinst d-i (testing), so there are only /dev/sda1 and the swap
partition present. I used ext4 as the file system.
I'm also having the drive checked by smartmontools at boot time and
have received no warnings.
You're basically not supposed to get I/O errors on drives like that. I'd
try running smartctl -a /dev/sda; or similar just to see whether any
errors have occured on the drive. It's possible that there's a bad
sector early on which is only exposed when something tries to find a gpt
partition table, or it could be a bug in systemd-gpt-auto-generator
which your particular setup is triggering.
You might be able to trigger it with gdisk -l /dev/sda; or similar, too.
I checked with smartctl and was told the test completed without error.
There were no errors in the log at all.
However, the output from the gdisk command was:
GPT fdisk (gdisk) version 0.8.10
Partition table scan:
MBR: MBR only
BSD: not present
APM: not present
GPT: not present
***************************************************************
Found invalid GPT and valid MBR; converting MBR to GPT format
in memory.
***************************************************************
Disk /dev/sda: 976773168 sectors, 465.8 GiB
Logical sector size: 512 bytes
Disk identifier (GUID): 43493272-B516-4A14-95E5-BF0E895243CB
Partition table holds up to 128 entries
First usable sector is 34, last usable sector is 976773134
Partitions will be aligned on 2048-sector boundaries
Total free space is 6125 sectors (3.0 MiB)
Number Start (sector) End (sector) Size Code Name
1 2048 943697919 450.0 GiB 8300 Linux filesystem
5 943699968 976771071 15.8 GiB 8200 Linux swap
I can see the words "invalid GPT and valid MBR" in that report, but --
save for the sizes and locations pertinent to the different disks --
this is exactly the same output that command gives me on my other systems.
Do you see anything significant?
If not, I'll try my hand at filing a severity minor bug against systemd
to see if the maintainers think anything of it.
If that doesn't turn up anything useful, file a bug against systemd, and
ask the maintainers what additional debugging information you can
provide. [It's probably severity minor, since this particularly failure
isn't going to hurt anything.]
I thought I ought to check to see if anyone thinks this is likely to
indicate that I'm about to get bit in the butt.
I'd make sure that I had my backups in order, but that's really just out
an abundance of caution.
Yup. I'm meticulous about backup strategy and practice. I'm retired, so
I have plenty of time to implement it. I never allow myself any excuses.
Thank you for your suggestions.
Jape
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/543ea384.6090...@comcast.net