<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>Hallo,</p>
    <p>I have strange problem: when I reset the node on which my
      resources are running, they are correctly migrated to the other
      node. But when I turn the failed node back, then as soon as it is
      up all resources are returned back to it. I have set
      resource-stickiness default value to 100. When this did not help I
      have set up resource-stickiness meta attr also to 100 for all my
      resources. Still when the failed node recovers the resources are
      migrated back to it! Where should I look to try to understand this
      situation?</p>
    <p>Here's the configuration of my cluster:<br>
    </p>
    root@node1# pcs status<br>
    Cluster name: gcluster<br>
    Cluster Summary:<br>
      * Stack: corosync<br>
      * Current DC: node1 (version 2.0.4-2deceaa3ae) - partition with
    quorum<br>
      * Last updated: Sat Sep 26 11:12:34 2020<br>
      * Last change:  Sat Sep 26 10:39:16 2020 by root via cibadmin on
    node1<br>
      * 2 nodes configured<br>
      * 14 resource instances configured (1 DISABLED)<br>
    <br>
    Node List:<br>
      * Online: [ node1 node2 ]<br>
    <br>
    Full List of Resources:<br>
      * ilo5_node1    (stonith:fence_ilo5_ssh):     Started node2<br>
      * ilo5_node2    (stonith:fence_ilo5_ssh):     Started node1<br>
      * Resource Group: VirtIP:<br>
        * PrimaryIP    (ocf::heartbeat:IPaddr2):     Started node2<br>
        * PrimaryIP6    (ocf::heartbeat:IPv6addr):     Started node2<br>
        * AliasIP    (ocf::heartbeat:IPaddr2):     Started node2<br>
      * BackupFS    (ocf::redhat:netfs.sh):     Started node2<br>
      * Clone Set: MailVolume-clone [MailVolume] (promotable):<br>
        * Masters: [ node2 ]<br>
        * Slaves: [ node1 ]<br>
      * MailFS    (ocf::heartbeat:Filesystem):     Started node2<br>
      * apache    (ocf::heartbeat:apache):     Started node2<br>
      * postfix    (ocf::heartbeat:postfix):     Started node2<br>
      * amavis    (service:amavis):     Started node2<br>
      * dovecot    (service:dovecot):     Started node2<br>
      * openvpn    (service:openvpn):     Stopped (disabled)<br>
    <br>
    And resources:<br>
    <br>
    root@node1# pcs resource config<br>
     Group: VirtIP<br>
      Meta Attrs: resource-stickiness=100<br>
      Resource: PrimaryIP (class=ocf provider=heartbeat type=IPaddr2)<br>
       Attributes: cidr_netmask=16 ip=xx.xx.xx.20 nic=br0<br>
       Meta Attrs: resource-stickiness=100<br>
       Operations: monitor interval=30s (PrimaryIP-monitor-interval-30s)<br>
                   start interval=0s timeout=20s
    (PrimaryIP-start-interval-0s)<br>
                   stop interval=0s timeout=20s
    (PrimaryIP-stop-interval-0s)<br>
      Resource: PrimaryIP6 (class=ocf provider=heartbeat type=IPv6addr)<br>
       Attributes: cidr_netmask=64 ipv6addr=xxxx:xxxx:xxxx:xxxx:0:0:0:20
    nic=br0<br>
       Meta Attrs: resource-stickiness=100<br>
       Operations: monitor interval=30s
    (PrimaryIP6-monitor-interval-30s)<br>
                   start interval=0s timeout=15s
    (PrimaryIP6-start-interval-0s)<br>
                   stop interval=0s timeout=15s
    (PrimaryIP6-stop-interval-0s)<br>
      Resource: AliasIP (class=ocf provider=heartbeat type=IPaddr2)<br>
       Attributes: cidr_netmask=16 ip=xx.xx.yy.20 nic=br0<br>
       Meta Attrs: resource-stickiness=100<br>
       Operations: monitor interval=30s (AliasIP-monitor-interval-30s)<br>
                   start interval=0s timeout=20s
    (AliasIP-start-interval-0s)<br>
                   stop interval=0s timeout=20s
    (AliasIP-stop-interval-0s)<br>
     Resource: BackupFS (class=ocf provider=redhat type=netfs.sh)<br>
      Attributes: export=/Backup/Gateway fstype=nfs host=atlas
    mountpoint=/Backup options=noatime,async<br>
      Meta Attrs: resource-stickiness=100<br>
      Operations: monitor interval=1m timeout=10
    (BackupFS-monitor-interval-1m)<br>
                  monitor interval=5m timeout=30 OCF_CHECK_LEVEL=10
    (BackupFS-monitor-interval-5m)<br>
                  monitor interval=10m timeout=30 OCF_CHECK_LEVEL=20
    (BackupFS-monitor-interval-10m)<br>
                  start interval=0s timeout=900
    (BackupFS-start-interval-0s)<br>
                  stop interval=0s timeout=30
    (BackupFS-stop-interval-0s)<br>
     Clone: MailVolume-clone<br>
      Meta Attrs: clone-max=2 clone-node-max=1 notify=true
    promotable=true promoted-max=1 promoted-node-max=1
    resource-stickiness=100<br>
      Resource: MailVolume (class=ocf provider=linbit type=drbd)<br>
       Attributes: drbd_resource=mail<br>
       Meta Attrs: resource-stickiness=100<br>
       Operations: demote interval=0s timeout=90
    (MailVolume-demote-interval-0s)<br>
                   monitor interval=60s
    (MailVolume-monitor-interval-60s)<br>
                   notify interval=0s timeout=90
    (MailVolume-notify-interval-0s)<br>
                   promote interval=0s timeout=90
    (MailVolume-promote-interval-0s)<br>
                   reload interval=0s timeout=30
    (MailVolume-reload-interval-0s)<br>
                   start interval=0s timeout=240
    (MailVolume-start-interval-0s)<br>
                   stop interval=0s timeout=100
    (MailVolume-stop-interval-0s)<br>
     Resource: MailFS (class=ocf provider=heartbeat type=Filesystem)<br>
      Attributes: device=/dev/drbd0 directory=/var/mail fstype=btrfs<br>
      Meta Attrs: resource-stickiness=100<br>
      Operations: monitor interval=20s timeout=40s
    (MailFS-monitor-interval-20s)<br>
                  start interval=0s timeout=60s
    (MailFS-start-interval-0s)<br>
                  stop interval=0s timeout=60s (MailFS-stop-interval-0s)<br>
     Resource: apache (class=ocf provider=heartbeat type=apache)<br>
      Attributes: client=wget statusurl=<a class="moz-txt-link-freetext" href="https://localhost/server-status">https://localhost/server-status</a><br>
      Meta Attrs: resource-stickiness=100<br>
      Operations: monitor interval=1min (apache-monitor-interval-1min)<br>
                  start interval=0s timeout=40s
    (apache-start-interval-0s)<br>
                  stop interval=0s timeout=60s (apache-stop-interval-0s)<br>
     Resource: postfix (class=ocf provider=heartbeat type=postfix)<br>
      Meta Attrs: resource-stickiness=100<br>
      Operations: monitor interval=60s timeout=20s
    (postfix-monitor-interval-60s)<br>
                  reload interval=0s timeout=20s
    (postfix-reload-interval-0s)<br>
                  start interval=0s timeout=20s
    (postfix-start-interval-0s)<br>
                  stop interval=0s timeout=20s
    (postfix-stop-interval-0s)<br>
     Resource: amavis (class=service type=amavis)<br>
      Meta Attrs: resource-stickiness=100<br>
      Operations: force-reload interval=0s timeout=15
    (amavis-force-reload-interval-0s)<br>
                  monitor interval=15 timeout=15
    (amavis-monitor-interval-15)<br>
                  restart interval=0s timeout=15
    (amavis-restart-interval-0s)<br>
                  start interval=0s timeout=15
    (amavis-start-interval-0s)<br>
                  stop interval=0s timeout=15 (amavis-stop-interval-0s)<br>
     Resource: dovecot (class=service type=dovecot)<br>
      Meta Attrs: resource-stickiness=100<br>
      Operations: force-reload interval=0s timeout=15
    (dovecot-force-reload-interval-0s)<br>
                  monitor interval=15 timeout=15
    (dovecot-monitor-interval-15)<br>
                  restart interval=0s timeout=15
    (dovecot-restart-interval-0s)<br>
                  start interval=0s timeout=15
    (dovecot-start-interval-0s)<br>
                  stop interval=0s timeout=15 (dovecot-stop-interval-0s)<br>
     Resource: openvpn (class=service type=openvpn)<br>
      Meta Attrs: resource-stickiness=100 target-role=Stopped<br>
      Operations: force-reload interval=0s timeout=15
    (openvpn-force-reload-interval-0s)<br>
                  monitor interval=15 timeout=15
    (openvpn-monitor-interval-15)<br>
                  restart interval=0s timeout=15
    (openvpn-restart-interval-0s)<br>
                  start interval=0s timeout=15
    (openvpn-start-interval-0s)<br>
                  stop interval=0s timeout=15 (openvpn-stop-interval-0s)<br>
    <p>drbd resource is configured as follows:</p>
    <p>root@node1# cat /etc/drbd.d/mail.res <br>
      resource mail {<br>
        protocol  B;<br>
        device    /dev/drbd0;<br>
        disk      /dev/sys/mail;<br>
        meta-disk internal;<br>
      <br>
        net {<br>
          csums-alg sha1;<br>
          after-sb-0pri discard-zero-changes;<br>
          after-sb-1pri discard-secondary;<br>
          after-sb-2pri disconnect;<br>
          rr-conflict disconnect;<br>
        }<br>
      <br>
        handlers {<br>
          fence-peer            "/usr/lib/drbd/crm-fence-peer.sh";<br>
          after-resync-target   "/usr/lib/drbd/crm-unfence-peer.sh";<br>
          split-brain           "/usr/lib/drbd/notify-split-brain.sh
      <a class="moz-txt-link-abbreviated" href="mailto:admin@logit-ag.de">admin@logit-ag.de</a>";<br>
        }<br>
      <br>
        on node1 {<br>
          address 192.168.0.102:7789;<br>
        }<br>
        on node2 {<br>
          address 192.168.0.103:7789;<br>
        }<br>
      }<br>
    </p>
    <p>Best regards,<br>
    </p>
    <pre class="moz-signature" cols="128">-- 
 \   / |                                   |
 (OvO) |  Mikhail Iwanow                   |
 (^^^) |                                   |
  \^/  |      E-mail:  <a class="moz-txt-link-abbreviated" href="mailto:ivans@logit-ag.de">ivans@logit-ag.de</a>   |
  ^ ^  |                                   |
</pre>
  </body>
</html>