HAProxy prechecks fail when scaling out to a new host with --limit or --serial

[Mark Goddard]

Steps to reproduce
------------------

Deploy HAProxy on one or more servers. Add another server to the inventory in the haproxy group, and run the following:

kolla-ansible prechecks --limit <new host>

Expected results
----------------

Prechecks pass

Actual results
--------------

TASK [haproxy : Checking if kolla_internal_vip_address and kolla_external_vip_address are not pingable from any node] **************************************************************************************
failed: [controller0] (item={u'command': u'ping', u'address': u'192.168.33.2'}) => {"ansible_loop_var": "item", "changed": false, "cmd": ["ping", "-c", "3", "192.168.33.2"], "delta": "0:00:02.059599", "end": "2020-03-24 17:59:56.080378", "failed_when_result": true, "item": {"address": "192.168.33.2", "command": "ping"}, "rc": 0, "start": "2020-03-24 17:59:54.020779", "stderr": "", "stderr_lines": [], "stdout": "PING 192.168.33.2 (192.168.33.2) 56(84) bytes of data.\n64 bytes from 192.168.33.2: icmp_seq=1 ttl=64 time=0.479 ms\n64 bytes from 192.168.33.2: icmp_seq=2 ttl=64 time=0.219 ms\n64 bytes from 192.168.33.2: icmp_seq=3 ttl=64 time=0.172 ms\n\n--- 192.168.33.2 ping statistics ---\n3 packets transmitted, 3 received, 0% packet loss, time 57ms\nrtt min/avg/max/mdev = 0.172/0.290/0.479/0.135 ms", "stdout_lines": ["PING 192.168.33.2 (192.168.33.2) 56(84) bytes of data.", "64 bytes from 192.168.33.2: icmp_seq=1 ttl=64 time=0.479 ms", "64 bytes from 192.168.33.2: icmp_seq=2 ttl=64 time=0.219 ms", "64 bytes from 192.168.33.2: icmp_seq=3 ttl=64 time=0.172 ms", "", "--- 192.168.33.2 ping statistics ---", "3 packets transmitted, 3 received, 0% packet loss, time 57ms", "rtt min/avg/max/mdev = 0.172/0.290/0.479/0.135 ms"]}
failed: [controller0] (item={u'command': u'ping', u'address': u'192.168.33.2'}) => {"ansible_loop_var": "item", "changed": false, "cmd": ["ping", "-c", "3", "192.168.33.2"], "delta": "0:00:02.032246", "end": "2020-03-24 17:59:58.576390", "failed_when_result": true, "item": {"address": "192.168.33.2", "command": "ping"}, "rc": 0, "start": "2020-03-24 17:59:56.544144", "stderr": "", "stderr_lines": [], "stdout": "PING 192.168.33.2 (192.168.33.2) 56(84) bytes of data.\n64 bytes from 192.168.33.2: icmp_seq=1 ttl=64 time=0.279 ms\n64 bytes from 192.168.33.2: icmp_seq=2 ttl=64 time=0.181 ms\n64 bytes from 192.168.33.2: icmp_seq=3 ttl=64 time=0.271 ms\n\n--- 192.168.33.2 ping statistics ---\n3 packets transmitted, 3 received, 0% packet loss, time 30ms\nrtt min/avg/max/mdev = 0.181/0.243/0.279/0.047 ms", "stdout_lines": ["PING 192.168.33.2 (192.168.33.2) 56(84) bytes of data.", "64 bytes from 192.168.33.2: icmp_seq=1 ttl=64 time=0.279 ms", "64 bytes from 192.168.33.2: icmp_seq=2 ttl=64 time=0.181 ms", "64 bytes from 192.168.33.2: icmp_seq=3 ttl=64 time=0.271 ms", "", "--- 192.168.33.2 ping statistics ---", "3 packets transmitted, 3 received, 0% packet loss, time 30ms", "rtt min/avg/max/mdev = 0.181/0.243/0.279/0.047 ms"]}

This happens because ansible does not execute on hosts where haproxy/keepalived is running, and therefore does not know that the VIP should be active.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.opendev.org/714950

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/714950
Committed: https://git.openstack.org/cgit/openstack/kolla-ansible/commit/?id=f3350d4e130b17b4035a8b67420d7015fc6cc99b
Submitter: Zuul
Branch: master

commit f3350d4e130b17b4035a8b67420d7015fc6cc99b
Author: Mark Goddard <email address hidden>
Date: Wed Mar 25 11:00:31 2020 +0000

    Fix HAProxy prechecks during scale-out with limit

    Deploy HAProxy on one or more servers. Add another server to the
    inventory in the haproxy group, and run the following:

    kolla-ansible prechecks --limit <new host>

    The following task will fail:

        TASK [haproxy : Checking if kolla_internal_vip_address and
        kolla_external_vip_address are not pingable from any node]

    This happens because ansible does not execute on hosts where
    haproxy/keepalived is running, and therefore does not know that the VIP
    should be active.

    This change skips VIP prechecks when not all HAProxy hosts are in the
    play.

    Closes-Bug: #1868986

    Change-Id: Ifbc73806b768f76f803ab01c115a9e5c2e2492ac

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/train
Review: https://review.opendev.org/716915

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/stein
Review: https://review.opendev.org/716917

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/716915
Committed: https://git.openstack.org/cgit/openstack/kolla-ansible/commit/?id=79747c4d8dfcad54a120d3f587501e8e7300d5b7
Submitter: Zuul
Branch: stable/train

commit 79747c4d8dfcad54a120d3f587501e8e7300d5b7
Author: Mark Goddard <email address hidden>
Date: Wed Mar 25 11:00:31 2020 +0000

    Fix HAProxy prechecks during scale-out with limit

    Deploy HAProxy on one or more servers. Add another server to the
    inventory in the haproxy group, and run the following:

    kolla-ansible prechecks --limit <new host>

    The following task will fail:

        TASK [haproxy : Checking if kolla_internal_vip_address and
        kolla_external_vip_address are not pingable from any node]

    This happens because ansible does not execute on hosts where
    haproxy/keepalived is running, and therefore does not know that the VIP
    should be active.

    This change skips VIP prechecks when not all HAProxy hosts are in the
    play.

    Closes-Bug: #1868986

    Change-Id: Ifbc73806b768f76f803ab01c115a9e5c2e2492ac
    (cherry picked from commit f3350d4e130b17b4035a8b67420d7015fc6cc99b)

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/716917
Committed: https://git.openstack.org/cgit/openstack/kolla-ansible/commit/?id=2e514a8cf310afc7e617813ab2ad1080eb72a575
Submitter: Zuul
Branch: stable/stein

commit 2e514a8cf310afc7e617813ab2ad1080eb72a575
Author: Mark Goddard <email address hidden>
Date: Wed Mar 25 11:00:31 2020 +0000

    Fix HAProxy prechecks during scale-out with limit

    Deploy HAProxy on one or more servers. Add another server to the
    inventory in the haproxy group, and run the following:

    kolla-ansible prechecks --limit <new host>

    The following task will fail:

        TASK [haproxy : Checking if kolla_internal_vip_address and
        kolla_external_vip_address are not pingable from any node]

    This happens because ansible does not execute on hosts where
    haproxy/keepalived is running, and therefore does not know that the VIP
    should be active.

    This change skips VIP prechecks when not all HAProxy hosts are in the
    play.

    Closes-Bug: #1868986

    Change-Id: Ifbc73806b768f76f803ab01c115a9e5c2e2492ac
    (cherry picked from commit f3350d4e130b17b4035a8b67420d7015fc6cc99b)
    (cherry picked from commit 79747c4d8dfcad54a120d3f587501e8e7300d5b7)