Arch just dies, no log info.

[inkydinky]

I'm running an up to date Arch install on a Beaglebone Black.
I run it headless and ssh in and use tmux to persist sessions.
I use the BBB as a file server and it has an external USB HDD attached.
Occasionally it will just die. The tmux session will die. I'll be able to switch between tabs in tmux but any keystrokes appear to do nothing. I believe at this point the BBB is dead in the water.

I've attempted to look at the logs but I only see `btmp btmp.1 faillog journal lastlog old pacman.log wtmp` as the contents of `/var/log`

Where should I be looking to begin to solve this problem?

Thanks.

[WarheadsSE]

logs are in the journal format.

How are you still able to move around tmux if no keystrokes are observed? That takes a keystroke, does it not?

[inkydinky]

To be more specific, I can switch tmux tabs but when I input a command nothing happens. A simple `ls` will just sit there with no output. When I Ctrl+C I just get the ^C to appear but never make it back to a prompt. I then go back and just pull the plug to reset the BBB.

If the BBB is in this state I can attempt to ssh in again but the ssh just hangs. It never gives me a pwd prompt.

[WarheadsSE]

So tmux is responding, but programs requiring disk interaction appear to be hanging. Would you consider leaving top or even htop running in a pane?

[inkydinky]

I will leave top open and see where that gets me.

I did a `journalctl -p err` and saw this:

Dec 31 17:39:39 alarm kernel: net eth0: phy 4a101000.mdio:01 not found on slave 1
-- Reboot --
Feb 18 06:14:40 alarm systemd-coredump[26530]: Process 292 (main) of user 1000 dumped core.
-- Reboot --
Feb 28 10:46:48 alarm systemd-coredump[329]: Process 322 (main) of user 1000 dumped core.
-- Reboot --
Mar 17 19:45:27 alarm sshd[365]: pam_tally(sshd:setcred): Tally underflowed for user beagleboneblack

beagleboneblack@alarm desks]$ journalctl --since=2015-02-18 --until=2015-02-19
-- Logs begin at Thu 2013-11-28 10:29:19 MST, end at Wed 2015-03-18 18:01:01 MDT. --
Feb 18 06:14:40 alarm systemd-coredump[26530]: Process 292 (main) of user 1000 dumped core.
Feb 18 06:26:46 alarm systemd[281]: Starting Shutdown.
Feb 18 06:26:46 alarm systemd[281]: Reached target Shutdown.
Feb 18 06:26:46 alarm systemd[281]: Starting Exit the Session...
Feb 18 06:26:46 alarm systemd[281]: Stopping Default.
Feb 18 06:26:46 alarm systemd[281]: Stopped target Default.

But that didn't get me anywhere and I didn't see a `core` to debug. I looked in home and /var/lib/systemd/coredump but didn't see a core.
I'll look into enabling core dumping and hopefully get more clues next crash.