1. Booting On LAN / filesystem check

• in order to boot pi2 computer from pi1, with system mounted by NFS ( Boot ON LAN ) do
  
  1.  login to pi1 as root ( su - )
  2. run :
       /etc/init.d/dhcpd start
       /etc/init.d/nfs start
     
              /etc/init.d/nfslock start
  1. login to pi2 as root and type:
       reboot
  2. possibly it is possible to login to pi2 only from pi1 machine
     
  3. check disks using command :
     
       e2fsck  /dev/hda1
       e2fsck  /dev/hda2
       e2fsck  /dev/hda5
       e2fsck  /dev/hdb3
       e2fsck  /dev/hdc3
     
       or  force checking even in case filesystem seems to be clean :
     
     
    e2fsck -f /dev/hda1
    e2fsck -f /dev/hda2
    e2fsck -f /dev/hda5
    e2fsck -f /dev/hdb3
    e2fsck -f /dev/hdc3
  
  Message similar to the following is expected, otherwise call experts :
      fsck 1.35 (28-Feb-2004)
      e2fsck 1.35 (28-Feb-2004)
      /: clean, 195820/1281696 files, 2163245/2560351 blocks (check in 2 mounts)
  
  
  in order to check for bad blocks addtional check can be run,
  BUT DO IT ONLY AFTER CONFIRMATION FROM
  PI-EXPERTS
  ( NOTE - can be very time consuming ) :
    
     fsck /dev/hda1 -cy
     ...
  
  
  1. after diagnostics is finished, on pi1 do :
       /etc/init.d/dhcpd stop
     
  2. and reboot pi2 machine to start from its disks :
      reboot
  3. or shutdown :
     shutdown -h now
     
     
• in order to boot pi1 computer from net do same actions but on oposite machines ( p1 <-> pi2 )
  
  

2. Starting system manaully (pi2):

• verify that night script was generated in /opt/pi/dev/pisys/scripts, if not run ( it never overwrites existing script ! ) , crontab at 19:00 CLT:
      autogen.sh
• check if processes : last2jpg! , update_alert_html.sh are running if not run ( crontab at  19:00 CLT ):
      start_monitor.sh
• check for disk space, possibly auto-cleaner was not executed ( crontab at 19:05 CLT ) , run :
     clean_online.sh 6
• to start system run ( crontab at 20:00 CLT ) :
    /opt/pi/dev/pisys/run_pisys! check_dome 28800
• check if SLT is running ( ps -ef | grep evtanal ) and if not run ( crontab at 20:30 CLT  ):
   /opt/pi/dev/pisys/daq/ndir/bin/runslt! >> /opt/pi/dev/pisys/log/slt.log 2>&1
• logging information from sensors, login as root and run ( crontab at 19:05 CLT ):
   log_sensors.sh
  

3. Current pi-system emergency procedure

•  pi2 is not running after morning shutdown
  
•  pi1 is running
  

• make sure dhcp server is not running on pi1 :
     /etc/init.d/dhcpd stop
  
• wake pi2 machine :
      wakepi2!
  NOTE : see point 4. ( before ) - automatic disk check can be executed ( every 10 mounts ) so this can take more time
  then you expect  - do not act to quickly !!!!
  
• verify in /var/log/messages that pi2 is up and by pinging it ( ping pi2 ) , login to pi2 to be sure
  
• after pi2 is up ( and you already logged into it ) , run dhcp server back ( in case pi2 reboots spontanously it will run from pi1 ) , enabling of autofs is very important !!! :
    /etc/init.d/dhcpd start
    /etc/init.d/nfs start
    /etc/init.d/nfslock start
    /etc/init.d/autofs start
  

•  check point 2.  - if some of this actions need to be taken : autogen.sh , start_monitor.sh , clean , but not run_pisys! and runslt! - this two should be
   executed from crontab
  
• 
  

• not ealier then 11:20
• login to pi1 as root and turn off nfs services :
     /etc/init.d/nfs stop
     /etc/init.d/nfslock stop
     /etc/init.d/autofs stop
  
• log in to pi2 as root and run :
    shutdown -h now
• then log in to pi1 machine as root and turn off dhcp server :
     /etc/init.d/dhcpd stop

4. Automatic disk diagnostic settings :

5. Pi2 usage log file