[conspire] debugging

[Paul Zander]

Yesterday there were several emails and live discussion at CABAL.


The concluding recommendation was that I should boot from live CD/USB and run fsck on the root partition. I had a vague recollection of being able to just force fsck on reboot.  A little web browsing found this:
    to force fsck on reboot 
    # shutdown -fF  now

Followed by the comment that it "many not work on many modern distros".  This explains why I have a distant memory that doesn't match the current man page.

The following site was helpful:
https://linuxconfig.org/how-to-force-fsck-to-check-filesystem-after-system-reboot-on-linux

Following the instructions to use tune2fs I got this:
    Last mounted on:          /
    Last checked:             Sat Aug  2 14:05:47 2014

That is a long time.

I followed more of the details on that page to force fsck during the reboot.  There were a few seconds with messages including, check forced ... clean
Now the root partition shows:
    Last checked:             Sun Feb 12 08:27:47 2017

Now that I was focused on the messages, I able to correctly capture the following error.  It was displayed for maybe 15 seconds before messages started scrolling.

What I transcribed:
    blk_update_request: I/O error, dev fd0, sector 0

from syslog:
    Feb 12 08:19:50 PZ01 kernel: [  915.536470] blk_update_request: I/O error, dev fd0, sector 0

Ah this computer doesn't even have a floppy drive. It is altogether possible that this message has been there forever without my noticing.

Going back a couple days, my attention "was got" when this was the last message followed by nothing but blinky disk activity LED for several minutes.


That really long "pause" hasn't repeated during several subsequent reboots.  If it doesn't repeat, there is nothing to fix.  Except for the missing grub symbol table, but that appears to be unrelated matter and not the direct cause of the hang.