4 Replies Latest reply: May 28, 2012 2:54 AM by 468439 RSS

    Problems starting crs after reboot

    468439
      I've made some progress getting Oracle RAC installed but after installing grid infrastructure I was distracted by other work that needed doing and when I came back one of my nodes had rebooted. It appears CRS was not working on that node after the reboot. I remembered having to do:

      /bin/dd if=/var/tmp/.oracle/npohasd of=/dev/null bs=1024 count=1

      whilst running the root.sh script (see root.sh failing on second node with "ora.asm -init" failed and thought this might be the problem again.

      It appears the only way I can get this node to startup correctly is to:

      1. rename /etc/init.d/ohasd to /etc/init.d/ohasd.saved
      2. reboot
      3. delete contents on /var/tmp/.oracle
      4.
      a. in session one run /etc/init.d/ohasd.saved start
      b. in session two wait for /var/tmp/.oracle/npohasd to appear and run the dd command above.

      This is Oracle Unbreakable Linux 6.2 and Oracle 11.2.

      Surely I don't have to do this every time I reboot?

      Thanks
        • 2. Re: Problems starting crs after reboot
          onedbguru
          MartinJEvans wrote:
          I've made some progress getting Oracle RAC installed but after installing grid infrastructure I was distracted by other work that needed doing and when I came back one of my nodes had rebooted. It appears CRS was not working on that node after the reboot. I remembered having to do:

          /bin/dd if=/var/tmp/.oracle/npohasd of=/dev/null bs=1024 count=1
          I certainly hope you did not execute that dd command.

          it should be
          /bin/dd if=/dev/null of=/var/tmp/.oracle/npohasd bs=1024 count=1

          overwrite npohasd with /dev/null


          >
          Surely I don't have to do this every time I reboot?
          There is a very good startup map at: http://docs.oracle.com/cd/E14072_01/rac.112/e10717/intro.htm. This link should be in every RAC DBA's "favorites" or print out the map and post in your cubicle.

          Go to the GRID_HOME and do a find on *.log and using this map (about 2/3 down the page), start looking in the log files (in order of execution) to determine why it is not starting.

          Make sure there is an entry for ohasd in /etc/inittab and make sure that the ASM devices are starting up with the correct permissions. I created a chkconfig-urable file that I use to ensure that the devices have the correct permissions BEFORE the ohasd starts up (somewhere before the S90 range.)

          I will leave the creation of this file as an exercise for the OP.

          >
          Thanks
          • 3. Re: Problems starting crs after reboot
            468439
            I'm fairly sure the dd command is correct because a) it was from a MOS article but I cannot find it now (I thought it was 1069182.1) b) without it ohas does not start c) it is all over these forums and the net.

            There is an entry in /etc/inittab:

            h1:35:respawn:/etc/init.d/init.ohasd run >/dev/null 2>&1 </dev/null

            however, a man on inittab says:


            The /etc/inittab file was the configuration file used by the original System V
            init(8) daemon.

            The Upstart init(8) daemon does not use this file, and instead reads its configura-
            tion from files in /etc/init. See init(5) for more details.

            This is Unbreakable Linux 6.2. Perhaps inittab no longer being used is the problem.
            • 4. Re: Problems starting crs after reboot
              468439
              Answering my own problem:

              The issue appears to be that Oracle Unbreakable Linux 6.2 does not use inittab:


              The /etc/inittab file was the configuration file used by the original System V init(8) daemon.
              The Upstart init(8) daemon does not use this file, and instead reads its configuration from files in
              /etc/init. See init(5) for more details.

              So I created /etc/init/oracle-ohasd.conf as follows:

              # Oracle OHASD startup

              start on runlevel [35]
              stop on runlevel [!35]
              respawn
              exec /etc/init.d/init.ohasd run >/dev/null 2>&1 </dev/null

              rebooted and all was fine.