lancelakey

#!/usr/bin/env bash
# 
# because sometimes a server sporadically acts funny and we're all like wtf.

# date and -+ separated list of general os, system, and network stuff.
echo -e "#!/usr/bin/env bash\n(date; top -n1; ps axfu; iostat; vmstat; df -h; lsof -n -i; netstat -ano; w; echo -e '\n-----------------\n\n') >> /var/log/wtf.log" > /root/wtf.sh
chmod +x /root/wtf.sh

# put it in the crontab.
echo '* * * * * /root/wtf.sh' >> /var/spool/cron/crontabs/root

lancelakey

This was a bit of a trick to get set up, so here it is for anyone who comes across it in search results.

ZNC User Account

• Set up the account as usual (note the username and password)
• Set the defualt buffer size to 0. Every time the bot reconnects after a crash, it would try to respond to the commands it hear or respond in the buffer, and might crash it again.
• Go ahead and define a network and channel for your bot. Note the network name.
• You should now see a "ghost" for your hubot connected.
• Look at the ZNC user list. The hubot user will show 1 network, but 0 clients.