<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    Hi Sunil,<br>
    Sorry for my late reply, i just had time today to start from scratch
    and test.<br>
    I rebuilt my environment(2 nodes connected to a SAN via
    iSCSI+multipath). I still have the issue that the heartbeat is
    active after I umount my ocfs2 volume. <br>
    /etc/init.d/o2cb stop<br>
    Stopping O2CB cluster CLUST: Failed<br>
    Unable to stop cluster as heartbeat region still active<br>
    <br>
    ocfs2_hb_ctl -I -d /dev/mapper/volgr1-lvol0<br>
    0C4AB55FE9314FA5A9F81652FDB9B22D: 1 refs<br>
    <br>
    After i manually kill the ref (ocfs2_hb_ctl -K -d
    /dev/mapper/volgr1-lvol0 ocfs2
    ) i can stop successfully o2cb. I can live with that but why doesn't
    it stop automatically? As i understand, hearbeat should be started
    and stopped once the volume gets mounted/umounted.<br>
    <br>
    br,<br>
    Laurentiu.<br>
    <br>
    On 10/19/2011 02:28, Sunil Mushran wrote:
    <blockquote cite="mid:4E9E0B92.8060104@oracle.com" type="cite">
      <meta content="text/html; charset=ISO-8859-1"
        http-equiv="Content-Type">
      Manual delete will only work if there are no references. In your
      case<br>
      there are references.<br>
      <br>
      You may want to start both nodes from scratch. Do not start/stop<br>
      heartbeat manually. Also, do not force-format.<br>
      <br>
      On 10/18/2011 03:54 PM, Laurentiu Gosu wrote:
      <blockquote cite="mid:4E9E03B7.4080603@easic.ro" type="cite">
        <meta content="text/html; charset=ISO-8859-1"
          http-equiv="Content-Type">
        OK, i rebooted one of the nodes(both had similar issues); . But
        something is still fishy.<br>
        - i mounted the device: mount -t ocfs2 /dev/volgr1/lvol0
        /mnt/tmp/<br>
        - i unmount it: umount /mnt/tmp/<br>
        - tried to stop o2cb:&nbsp; /etc/init.d/o2cb stop<br>
        Stopping O2CB cluster CLUSTER: Failed<br>
        Unable to stop cluster as heartbeat region still active<br>
        - ocfs2_hb_ctl -I -u 0C4AB55FE9314FA5A9F81652FDB9B22D<br>
        0C4AB55FE9314FA5A9F81652FDB9B22D: 1 refs<br>
        -&nbsp; ocfs2_hb_ctl -K -u 0C4AB55FE9314FA5A9F81652FDB9B22D<br>
        ocfs2_hb_ctl: File not found by ocfs2_lookup while stopping
        heartbeat<br>
        - ls -Rl /sys/kernel/config/cluster/CLUSTER/heartbeat/<br>
        /sys/kernel/config/cluster/CLUSTER/heartbeat/:<br>
        total 0<br>
        drwxr-xr-x 2 root root&nbsp;&nbsp;&nbsp; 0 Oct 19 01:50
        0C4AB55FE9314FA5A9F81652FDB9B22D<br>
        -rw-r--r-- 1 root root 4096 Oct 19 01:40 dead_threshold<br>
        <br>
/sys/kernel/config/cluster/CLUSTER/heartbeat/0C4AB55FE9314FA5A9F81652FDB9B22D:<br>
        total 0<br>
        -rw-r--r-- 1 root root 4096 Oct 19 01:50 block_bytes<br>
        -rw-r--r-- 1 root root 4096 Oct 19 01:50 blocks<br>
        -rw-r--r-- 1 root root 4096 Oct 19 01:50 dev<br>
        -r--r--r-- 1 root root 4096 Oct 19 01:50 pid<br>
        -rw-r--r-- 1 root root 4096 Oct 19 01:50 start_block<br>
        <br>
        - i cannot manually delete
/sys/kernel/config/cluster/CLUSTER/heartbeat/0C4AB55FE9314FA5A9F81652FDB9B22D/<br>
        <br>
        PS: i'm going to sleep now, i have to be up in a few hours. We
        can continue tomorrow if it's ok with you. <br>
        Thank you for your help.<br>
        <br>
        Laurentiu.<br>
        <br>
        On 10/19/2011 01:33, Sunil Mushran wrote:
        <blockquote cite="mid:4E9DFEB0.9010206@oracle.com" type="cite">
          <meta content="text/html; charset=ISO-8859-1"
            http-equiv="Content-Type">
          One way this can happen is if one starts the hb manually and
          then force<br>
          formats on that volume. The format will generate a new uuid.
          Once that<br>
          happens, the hb tool cannot map the region to the device and
          thus fail<br>
          to stop it. Right now the easiest option on this box is
          resetting it.<br>
          <br>
          On 10/18/2011 03:24 PM, Laurentiu Gosu wrote:
          <blockquote cite="mid:4E9DFC93.1050109@easic.ro" type="cite">
            <meta content="text/html; charset=ISO-8859-1"
              http-equiv="Content-Type">
            Yes, i did reformat it(even more than once i think, last
            week). This is a pre-production system and i'm trying
            various options before moving into real life.<br>
            <br>
            <br>
            On 10/19/2011 01:19, Sunil Mushran wrote:
            <blockquote cite="mid:4E9DFB83.40603@oracle.com" type="cite">
              <meta content="text/html; charset=ISO-8859-1"
                http-equiv="Content-Type">
              Did you reformat the volume recently? or, when did you
              format last?<br>
              <br>
              On 10/18/2011 03:13 PM, Laurentiu Gosu wrote:
              <blockquote cite="mid:4E9DFA03.8030405@easic.ro"
                type="cite">
                <meta content="text/html; charset=ISO-8859-1"
                  http-equiv="Content-Type">
                well..this is weird<br>
                ls /sys/kernel/config/cluster/CLUSTER/heartbeat/<br>
                <b>918673F06F8F4ED188DDCE14F39945F6</b>&nbsp; dead_threshold<br>
                <br>
                looks like we have different UUIDs. Where is this coming
                from??<br>
                <br>
                ocfs2_hb_ctl -I -u 918673F06F8F4ED188DDCE14F39945F6<br>
                918673F06F8F4ED188DDCE14F39945F6: 1 refs<br>
                <br>
                <br>
                On 10/19/2011 01:04, Sunil Mushran wrote:
                <blockquote cite="mid:4E9DF7D0.7090404@oracle.com"
                  type="cite">Let's do it by hand. <br>
                  rm -rf /sys/kernel/config/cluster/.../heartbeat/<b>0C4AB55FE9314FA5A9F81652FDB9B22D






                  </b><br>
                  <br>
                  On 10/18/2011 02:52 PM, Laurentiu Gosu wrote: <br>
                  <blockquote type="cite">&nbsp;ocfs2_hb_ctl -K -u
                    0C4AB55FE9314FA5A9F81652FDB9B22D <br>
                    ocfs2_hb_ctl: File not found by ocfs2_lookup while
                    stopping heartbeat <br>
                    <br>
                    No improvment :( <br>
                    <br>
                    <br>
                    On 10/19/2011 00:50, Sunil Mushran wrote: <br>
                    <blockquote type="cite">See if this cleans it up. <br>
                      ocfs2_hb_ctl -K -u
                      0C4AB55FE9314FA5A9F81652FDB9B22D <br>
                      <br>
                      On 10/18/2011 02:44 PM, Laurentiu Gosu wrote: <br>
                      <blockquote type="cite">ocfs2_hb_ctl -I -u
                        0C4AB55FE9314FA5A9F81652FDB9B22D <br>
                        0C4AB55FE9314FA5A9F81652FDB9B22D: 0 refs <br>
                        <br>
                        <br>
                        On 10/19/2011 00:43, Sunil Mushran wrote: <br>
                        <blockquote type="cite">ocfs2_hb_ctl -l -u
                          0C4AB55FE9314FA5A9F81652FDB9B22D <br>
                          <br>
                          On 10/18/2011 02:40 PM, Laurentiu Gosu wrote:
                          <br>
                          <blockquote type="cite">mounted.ocfs2 -d <br>
                            Device&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; FS&nbsp;&nbsp;&nbsp;&nbsp; Stack&nbsp;
                            UUID&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Label <br>
                            /dev/mapper/volgr1-lvol0&nbsp; ocfs2&nbsp; o2cb&nbsp;&nbsp;
                            0C4AB55FE9314FA5A9F81652FDB9B22D&nbsp; ocfs2 <br>
                            <br>
                            mounted.ocfs2 -f <br>
                            Device&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; FS&nbsp;&nbsp;&nbsp;&nbsp; Nodes <br>
                            /dev/mapper/volgr1-lvol0&nbsp; ocfs2&nbsp; ro02xsrv001
                            <br>
                            <br>
                            ro02xsrv001 = the other node in the cluster.
                            <br>
                            <br>
                            By the way, there is no /dev/md-2 <br>
                            &nbsp;ls /dev/dm-* <br>
                            /dev/dm-0&nbsp; /dev/dm-1 <br>
                            <br>
                            <br>
                            On 10/19/2011 00:37, Sunil Mushran wrote: <br>
                            <blockquote type="cite">So it is not
                              mounted. But we still have a hb thread
                              because <br>
                              hb could not be stopped during umount. The
                              reason for that <br>
                              could be the same that causes ocfs2_hb_ctl
                              to fail. <br>
                              <br>
                              Do: <br>
                              mounted.ocfs2 -d <br>
                              <br>
                              On 10/18/2011 02:32 PM, Laurentiu Gosu
                              wrote: <br>
                              <blockquote type="cite">ls -lR
                                /sys/kernel/debug/ocfs2 <br>
                                /sys/kernel/debug/ocfs2: <br>
                                total 0 <br>
                                <br>
                                ls -lR /sys/kernel/debug/o2dlm <br>
                                /sys/kernel/debug/o2dlm: <br>
                                total 0 <br>
                                <br>
                                ocfs2_hb_ctl -I -d /dev/dm-2 <br>
                                ocfs2_hb_ctl: Device name specified was
                                not found while reading uuid <br>
                                <br>
                                There is no /dev/dm-2 mounted. <br>
                                <br>
                                <br>
                                On 10/19/2011 00:27, Sunil Mushran
                                wrote: <br>
                                <blockquote type="cite">mount -t debugfs
                                  debugfs /sys/kernel/debug <br>
                                  <br>
                                  Then list that dir. <br>
                                  <br>
                                  Also, do: <br>
                                  ocfs2_hb_ctl -l -d /dev/dm-2 <br>
                                  <br>
                                  Be careful before killing. We want to
                                  be sure that dev is not mounted. <br>
                                  <br>
                                  On 10/18/2011 02:23 PM, Laurentiu Gosu
                                  wrote: <br>
                                  <blockquote type="cite">Again&nbsp;&nbsp; the
                                    outputs: <br>
                                    &nbsp;cat
/sys/kernel/config/cluster/CLUSTER/heartbeat/918673F06F8F4ED188DDCE14F39945F6/dev<br>
                                    dm-2 <br>
                                    ---&gt;here should be volgr1-lvol0 i
                                    guess? <br>
                                    <br>
                                    ls -lR /sys/kernel/debug/ocfs2 <br>
                                    ls: /sys/kernel/debug/ocfs2: No such
                                    file or directory <br>
                                    <br>
                                    ls -lR /sys/kernel/debug/o2dlm <br>
                                    ls: /sys/kernel/debug/o2dlm: No such
                                    file or directory <br>
                                    <br>
                                    I think i have to enable debug first
                                    somehow..? <br>
                                    <br>
                                    Laurentiu. <br>
                                    <br>
                                    On 10/19/2011 00:17, Sunil Mushran
                                    wrote: <br>
                                    <blockquote type="cite">What does
                                      this return? <br>
                                      cat
/sys/kernel/config/cluster/CLUSTER/heartbeat/918673F06F8F4ED188DDCE14F39945F6/dev<br>
                                      <br>
                                      Also, do: <br>
                                      ls -lR /sys/kernel/debug/ocfs2 <br>
                                      ls -lR /sys/kernel/debug/o2dlm <br>
                                      <br>
                                      On 10/18/2011 02:14 PM, Laurentiu
                                      Gosu wrote: <br>
                                      <blockquote type="cite">Here is
                                        the output: <br>
                                        <br>
                                        ls -lR
                                        /sys/kernel/config/cluster <br>
                                        /sys/kernel/config/cluster: <br>
                                        total 0 <br>
                                        drwxr-xr-x 4 root root 0 Oct 19
                                        00:12 CLUSTER <br>
                                        <br>
                                        /sys/kernel/config/cluster/CLUSTER:


                                        <br>
                                        total 0 <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 fence_method <br>
                                        drwxr-xr-x 3 root root&nbsp;&nbsp;&nbsp; 0 Oct
                                        19 00:12 heartbeat <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 idle_timeout_ms <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 keepalive_delay_ms <br>
                                        drwxr-xr-x 4 root root&nbsp;&nbsp;&nbsp; 0 Oct
                                        11 20:23 node <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 reconnect_delay_ms <br>
                                        <br>
                                        /sys/kernel/config/cluster/CLUSTER/heartbeat:






                                        <br>
                                        total 0 <br>
                                        drwxr-xr-x 2 root root&nbsp;&nbsp;&nbsp; 0 Oct
                                        19 00:12
                                        918673F06F8F4ED188DDCE14F39945F6
                                        <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 dead_threshold <br>
                                        <br>
/sys/kernel/config/cluster/CLUSTER/heartbeat/<b>918673F06F8F4ED188DDCE14F39945F6</b>:
                                        <br>
                                        total 0 <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 block_bytes <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 blocks <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 dev <br>
                                        -r--r--r-- 1 root root 4096 Oct
                                        19 00:12 pid <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 start_block <br>
                                        <br>
                                        /sys/kernel/config/cluster/CLUSTER/node:




                                        <br>
                                        total 0 <br>
                                        drwxr-xr-x 2 root root 0 Oct 19
                                        00:12 ro02xsrv001 <br>
                                        drwxr-xr-x 2 root root 0 Oct 19
                                        00:12 ro02xsrv002 <br>
                                        <br>
                                        /sys/kernel/config/cluster/CLUSTER/node/ro02xsrv001:






                                        <br>
                                        total 0 <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 ipv4_address <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 ipv4_port <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 local <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 num <br>
                                        <br>
                                        /sys/kernel/config/cluster/CLUSTER/node/ro02xsrv002:






                                        <br>
                                        total 0 <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 ipv4_address <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 ipv4_port <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 local <br>
                                        -rw-r--r-- 1 root root 4096 Oct
                                        19 00:12 num <br>
                                        <br>
                                        <br>
                                        <br>
                                        <br>
                                        On 10/19/2011 00:12, Sunil
                                        Mushran wrote: <br>
                                        <blockquote type="cite">ls -lR
                                          /sys/kernel/config/cluster <br>
                                          <br>
                                          What does this return? <br>
                                          <br>
                                          On 10/18/2011 02:05 PM,
                                          Laurentiu Gosu wrote: <br>
                                          <blockquote type="cite">Hi, <br>
                                            I have a 2 nodes ocfs2
                                            cluster running UEK
                                            2.6.32-100.0.19.el5, <br>
                                            ocfs2console-1.6.3-2.el5,
                                            ocfs2-tools-1.6.3-2.el5. <br>
                                            My problem is that all the
                                            time when i try to run
                                            /etc/init.d/o2cb stop <br>
                                            it fails with this error: <br>
                                            &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Stopping O2CB cluster
                                            CLUSTER: Failed <br>
                                            &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Unable to stop cluster
                                            as heartbeat region still
                                            active <br>
                                            There is no active mount
                                            point. I tried to manually
                                            stop the heartdbeat <br>
                                            with "ocfs2_hb_ctl -K -d
                                            /dev/mapper/volgr1-lvol0
                                            ocfs2" (after finding <br>
                                            the refs number with
                                            "ocfs2_hb_ctl -I -d
                                            /dev/mapper/volgr1-lvol0 ").
                                            <br>
                                            But even if refs number is
                                            set to zero the "heartbeat
                                            region still <br>
                                            active" occurs. <br>
                                            How can i fix this? <br>
                                            <br>
                                            Thank you in advance. <br>
                                            Laurentiu. <br>
                                            <br>
                                            <br>
                                            _______________________________________________






                                            <br>
                                            Ocfs2-users mailing list <br>
                                            <a moz-do-not-send="true"
                                              class="moz-txt-link-abbreviated"
href="mailto:Ocfs2-users@oss.oracle.com">Ocfs2-users@oss.oracle.com</a>
                                            <br>
                                            <a moz-do-not-send="true"
                                              class="moz-txt-link-freetext"
href="http://oss.oracle.com/mailman/listinfo/ocfs2-users">http://oss.oracle.com/mailman/listinfo/ocfs2-users</a>
                                            <br>
                                          </blockquote>
                                          <br>
                                        </blockquote>
                                        <br>
                                      </blockquote>
                                      <br>
                                    </blockquote>
                                    <br>
                                  </blockquote>
                                  <br>
                                </blockquote>
                                <br>
                              </blockquote>
                              <br>
                            </blockquote>
                            <br>
                          </blockquote>
                          <br>
                        </blockquote>
                        <br>
                      </blockquote>
                      <br>
                    </blockquote>
                    <br>
                  </blockquote>
                  <br>
                </blockquote>
                <br>
              </blockquote>
              <br>
            </blockquote>
            <br>
          </blockquote>
          <br>
        </blockquote>
        <br>
      </blockquote>
      <br>
    </blockquote>
    <br>
  </body>
</html>