Restarting Farm XXx

If you need to restart ER Farm - do not forget to save runlist and later upload it to the farm again. This could be done with
../bin/$BFARCH/OprCmd.pl -iFarm,User,Debug -nbbr-lock stoperfeeder
../bin/$BFARCH/OprCmd.pl -iFarm,User,Debug -noprx saverunlist ERx ~/runlists/runlist.ERx.<date>.txt

restart farm as described below and then
../bin/$BFARCH/OprCmd.pl -iFarm,User,Debug -noprx submitrunlist ERx ~/runlists/runlist.ERx.<date>.txt
../bin/$BFARCH/OprCmd.pl -iFarm,User,Debug -ncls0 starterfeeder

This is not necessary for PC farms.


Only reaping and restarting:
  1. cd workx
  2. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x farminfo XXx
    Make sure the farm is asleep before you reap it, or you have a very good reason to reap it in the middle of a run. REMEMBER: reaping the farm only means that nothing new will be started and what is running now will no longer be monitored, reaping does NOT kill Elven etc!
  3. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x map
    You should see all the services running.
  4. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x systemreaper FarmControl
    You are asked if you are sure and if you saved the runlist... you should decide the first and if the second is important in the current situation.
  5. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x map
    You should not see any services running (the last message should be "1 LPF found."). If you do, do systemreaper again.
  6. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x startsystem FarmControl
    Make SURE you do NOT do this command twice, since, if you do, all the services will be started twice and the only way out is to again reap and then scratch the farm since in this case, usually reaping does not clean up everything...
  7. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x map
    Don't go on until the output from map looks something like this
    It might take about one minute...
  8. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x setshifter XXx your#
    If you don't know your# try ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x getshifers XXx
    The first phone number corresponds to 0, when the farm is restarted it defaults to this number.
    Thus, if you skip this, your fellow shifter (the one whose beeper number is ('0') will hate you.
  9. Submit your runs or whatever
If you also want to reload the configuration for the farm (if you just changed it etc...) without scratching the farm you can add the following steps after step 5, i.e. just before you restart the farm:
  1. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x cleanupconf
  2. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x loadconf /nfs/oprserv01/u1/XXx/prod/OprProcessingSys/OprConfXXx.xml


Reaping, scratching and restarting:
  1. cd workx
  2. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x farminfo XXx
    Make sure the farm is asleep before you reap it, or you have a very good reason to reap it in the middle of a run. REMEMBER: reaping the farm only means that nothing new will be started and what is running now will no longer be monitored, reaping does NOT kill Elven etc!
  3. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x systemreaper FarmControl
    You are asked if you are sure and if you saved the runlist... you should decide the first and if the second is important in the current situation.
  4. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x map
    You should not see any services running (the last message should be "1 LPF found."). If you do, do systemreaper again.
  5. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x scratchFarm XXx
    Warning: Can't be done as babaropr
  6. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x loadconf /nfs/oprserv01/u1/XXx/prod/OprProcessingSys/OprConfXXx.xml
  7. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x startsystem FarmControl
    Make SURE you do NOT do this command twice, since, if you do, all the services will be started twice and the only way out is to again reap and then scratch the farm since in this case, usually reaping does not clean up everything...
  8. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x map
    Don't go on until the output from map looks something like this
    It might take about one minute...
  9. ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x setshifter XXx your#
    If you don't know your# try ../bin/$BFARCH/OprCmd.pl -iUser,Farm -noprserv0x getshifers XXx
    The first phone number corresponds to 0, when the farm is restarted it defaults to this number.
    Thus, if you skip this, your fellow shifter (the one whose beeper number is ('0') will hate you.
  10. Submit your runs or whatever

Ada Elizabeth Rubin
Last modified: Thu Apr 21 10:51:30 CEST 2005