Fuel for OpenStack

[octane] Upgrade second and third controllers separately fails at "installing openstack" part

Bug #1584656 reported by Andrey Volochay on 2016-05-23

This bug affects 1 person

Affects		Status	Importance	Assigned to	Milestone
	Fuel for OpenStack	Fix Committed	Medium	Oleg S. Gelbukh	Fuel for OpenStack 8.0-updates

Bug Description

Detailed bug description:
During upgrade MOS6.0 to MOS8.0 via Octane(stable/8.0), I have faced with following bug.

By doing second and third controllers upgrade via Octane (octane upgrade-node <target_node_id> <seed_env_id>), deployment failed at "installing openstack" part.

Could not evaluate: Primitive 'p_dns' was not found in CIB!

Steps to reproduce:
1) Have MOS6.0 environment
2) Upgrade master node (fuel 6.0) to 6.1 and then to 7.0
3) Backup environment settings, following documented process
4) Prepare new master node (fuel 8.0)
5) Restore the backup from third step
5) Upgrade restored environment via Octane
6) Upgrade primary controller via Octane
7) Upgrade DB via Octane
8) Upgrade Control Plane via Octane
9) Upgrade second controller
10) Upgrade third controller

Expected results:
Upgrade of second and third controllers is successfully done

Actual result:
Puppet apply returns an error which in turn breaks node upgrade.

Reproducibility:
When you upgrade controllers one by one, but not the both at one time.

Workaround:
Add records about second and third controllers in corosync.conf at primary-controller before their upgrade.

Description of the environment:
Versions of components: MOS8.0, Octane(stable/8.0)
Network model: Environment uses network bonding
Related projects installed: Octane

Additional information:
it happens in octane/commands/upgrade_node.py

if isolated or len(nodes) == 1:
env_util.deploy_nodes(env, nodes)
else:
env_util.deploy_changes(env, nodes)

It's unsafe to manage controllers upgrade via count of nodes that are upgraded. Node's role should be the criterion.

See original description

Tags:

Andrey Volochay (wallavv) on 2016-05-23

description:

updated

Andrey Volochay (wallavv) on 2016-05-23

description:

updated

Oleksiy Molchanov (omolchanov) on 2016-05-23

Changed in fuel:
milestone:	none → 8.0-updates
assignee:	nobody → Fuel Octane (fuel-octane-team)
importance:	Undecided → High
status:	New → Confirmed
tags:	added: team-upgrades

Revision history for this message

Oleg S. Gelbukh (gelbuhos) wrote on 2016-05-25:

Upgrade of secondary controllers one by one never worked. I think it is even reflected in the docs somewhere. We could extend the validations to exclude this possibility, I think.

Oleg S. Gelbukh (gelbuhos) on 2016-05-25

Changed in fuel:
importance:	High → Medium

Revision history for this message

OpenStack Infra (hudson-openstack) wrote on 2016-05-26: Fix proposed to fuel-octane (master)

Fix proposed to branch: master
Review: https://review.openstack.org/321599

Changed in fuel:
assignee:	Fuel Octane (fuel-octane-team) → Oleg S. Gelbukh (gelbuhos)
status:	Confirmed → In Progress

Oleg S. Gelbukh (gelbuhos) on 2016-05-31

summary:

- [octane] Upgrade second and third controllers fails at "installing
- openstack" part
+ [octane] Upgrade second and third controllers separately fails at
+ "installing openstack" part

Revision history for this message

OpenStack Infra (hudson-openstack) wrote on 2016-06-22: Fix merged to fuel-octane (master)

Reviewed: https://review.openstack.org/321599
Committed: https://git.openstack.org/cgit/openstack/fuel-octane/commit/?id=6111d9b9de10b583c5a1699edcd0994e55a739e3
Submitter: Jenkins
Branch: master

commit 6111d9b9de10b583c5a1699edcd0994e55a739e3
Author: Oleg Gelbukh <email address hidden>
Date: Thu May 26 14:54:41 2016 +0000

Validate nodes roles to exclude upgrade of single controller

The proper sequence to upgrade controllers is:

    - upgrade primary controller in isolation
    - cut over to the new control plane
    - upgrade remaining controllers at once

    To exclude the possibility to upgrade single controller after
    the cutover, add validation of roles of nodes to upgrade and
    stop if the only node passed to the script is a CIC.

Change-Id: Ie91ae35acb7e36103004a59c92cb495837e0dca0
Closes-bug: 1584656

Changed in fuel:
status:	In Progress → Fix Committed

Revision history for this message

OpenStack Infra (hudson-openstack) wrote on 2016-06-22: Fix proposed to fuel-octane (stable/mitaka)

Fix proposed to branch: stable/mitaka
Review: https://review.openstack.org/332931

Revision history for this message

OpenStack Infra (hudson-openstack) wrote on 2016-06-22: Fix merged to fuel-octane (stable/mitaka)

Download full text (40.5 KiB)

Reviewed: https://review.openstack.org/332931
Committed: https://git.openstack.org/cgit/openstack/fuel-octane/commit/?id=24f3c396612eb0c69fbf06bab3cebbb9ca829426
Submitter: Jenkins
Branch: stable/mitaka

commit b34d749f1c281dafbdbf155bd86830dc8f2a9aa2
Author: Ilya Kharin <email address hidden>
Date: Wed Jun 22 21:10:56 2016 +0300

Support mock==1.8.0 in unit tests

Use assert_called_once_with instead of assert_called_once that was
introduced in 2.0.0.

Change-Id: Ifb7699d4b552d148984961727355f0e23b487c7c

commit d60f1905143bb7576ffd670595de1c4aeafd7b34
Author: Ilya Kharin <email address hidden>
Date: Wed Jun 22 02:01:52 2016 +0300

Allow to authorize by the predefined admin_token

The admin_token_auth middleware is added to keystone pipelines to
allow authorization by admin_token.

Change-Id: Ic03150305a669fad1446436a68051fb9aa25b892

commit cc7fab59f44ffef60285f8732d798f52469b2530
Author: Ilya Kharin <email address hidden>
Date: Wed Jun 22 01:47:01 2016 +0300

Reset default_domain_id before the keystone task

    The default_domain_id should be removed from keystone.conf after restore
    of DB and before to apply the keystone puppet task to avoid of using the
    configured domain as a default.

Change-Id: I05a6c48532e8042496b3d8ccef53d65bf8c44653

commit e3f82399d567dbcfc1ae9a1ecbddba7bf5028fc8
Author: Ilya Kharin <email address hidden>
Date: Wed Jun 22 01:32:47 2016 +0300

Add helper function to iterate over parameters

    The helpers.iterate_parameters function allows to iterate over lines of
    INI-like files along with a context of information, such section,
    parameter and value.

Change-Id: I55b179118116fd5dacf100754057ea6589782dc2

commit 5ed370a4fbb3369d0f75873d205ac7f0c655f93a
Author: Ilya Kharin <email address hidden>
Date: Wed Jun 22 01:18:05 2016 +0300

Add update_file context manager for local files

    The subprocess.update_file function provides an ability to update
    content of a local file by iterating over lines of an original file and
    forming a result content in a temporary file to replace the original
    file in the end. This function is very useful to change configuration
    files.

Change-Id: I433a5da67887b231400dd63131799019f45c277c

commit 58f31e6c5f408630c42565ee53de6b59457bc84c
Author: Oleg Gelbukh <email address hidden>
Date: Wed Jun 22 14:34:03 2016 +0000

Escape passwords passed to openstack client

    If special symbols are used in password for 'admin' user in
    OpenStack, octane passes them to command line client as is
    and it breaks shell.

Properly escape the password before passing it to subprocess.

Change-Id: Iad635aec6d5b5cc32975937e00205b7e89dc99d9
Closes-bug: 1585960

commit 896aba1191eeb59cf85cc8be6a2ae67e08b76070
Author: Yuriy Taraday <email address hidden>
Date: Wed Jun 22 15:28:45 2016 +0300

Add absolute_import to util/docker.py to avoid local tempfile module

Closes-Bug: 1595156
Change-Id: I9484efce6fa7aec1b41cf592f9e9768d85931fa7

commit 46586a62df962b2ce00b5d8a63f6fd34c920a1f4
Author: Oleg Gelbukh <email address hidden>
Date: Thu M...

Reviewed:  https://review.openstack.org/332931
Committed: https://git.openstack.org/cgit/openstack/fuel-octane/commit/?id=24f3c396612eb0c69fbf06bab3cebbb9ca829426
Submitter: Jenkins
Branch:    stable/mitaka

commit b34d749f1c281dafbdbf155bd86830dc8f2a9aa2
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Jun 22 21:10:56 2016 +0300

Support mock==1.8.0 in unit tests
    
    Use assert_called_once_with instead of assert_called_once that was
    introduced in 2.0.0.
    
    Change-Id: Ifb7699d4b552d148984961727355f0e23b487c7c

commit d60f1905143bb7576ffd670595de1c4aeafd7b34
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Jun 22 02:01:52 2016 +0300

Allow to authorize by the predefined admin_token
    
    The admin_token_auth middleware is added to keystone pipelines to
    allow authorization by admin_token.
    
    Change-Id: Ic03150305a669fad1446436a68051fb9aa25b892

commit cc7fab59f44ffef60285f8732d798f52469b2530
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Jun 22 01:47:01 2016 +0300

Reset default_domain_id before the keystone task
    
    The default_domain_id should be removed from keystone.conf after restore
    of DB and before to apply the keystone puppet task to avoid of using the
    configured domain as a default.
    
    Change-Id: I05a6c48532e8042496b3d8ccef53d65bf8c44653

commit e3f82399d567dbcfc1ae9a1ecbddba7bf5028fc8
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Jun 22 01:32:47 2016 +0300

Add helper function to iterate over parameters
    
    The helpers.iterate_parameters function allows to iterate over lines of
    INI-like files along with a context of information, such section,
    parameter and value.
    
    Change-Id: I55b179118116fd5dacf100754057ea6589782dc2

commit 5ed370a4fbb3369d0f75873d205ac7f0c655f93a
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Jun 22 01:18:05 2016 +0300

Add update_file context manager for local files
    
    The subprocess.update_file function provides an ability to update
    content of a local file by iterating over lines of an original file and
    forming a result content in a temporary file to replace the original
    file in the end. This function is very useful to change configuration
    files.
    
    Change-Id: I433a5da67887b231400dd63131799019f45c277c

commit 58f31e6c5f408630c42565ee53de6b59457bc84c
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Wed Jun 22 14:34:03 2016 +0000

Escape passwords passed to openstack client
    
    If special symbols are used in password for 'admin' user in
    OpenStack, octane passes them to command line client as is
    and it breaks shell.
    
    Properly escape the password before passing it to subprocess.
    
    Change-Id: Iad635aec6d5b5cc32975937e00205b7e89dc99d9
    Closes-bug: 1585960

commit 896aba1191eeb59cf85cc8be6a2ae67e08b76070
Author: Yuriy Taraday <yorik.sar@gmail.com>
Date:   Wed Jun 22 15:28:45 2016 +0300

Add absolute_import to util/docker.py to avoid local tempfile module
    
    Closes-Bug: 1595156
    Change-Id: I9484efce6fa7aec1b41cf592f9e9768d85931fa7

commit 46586a62df962b2ce00b5d8a63f6fd34c920a1f4
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu May 26 13:07:52 2016 +0000

Improve partitions and disk metadata handling
    
    Partitions schema and disk metadata at nodes must be updated only
    if they are scheduled for reprovisioning, and the provisioning
    method has changed to incompatible (i.e. from 'cobbler' to 'image')
    due to the shift of release version.
    
    Move logic that manages partitions from upgrade handlers to function
    that actually initiates the re-provisioning of node. This ensures
    that for every node this is applied only once and only if necessary.
    
    Change-Id: I28718564543e3ccd92d985bd79c76e509e8eae1f
    Closes-bug: 1585938
    Closes-bug: 1575054

commit b5d884296c35c5bf41ec3928c7a96d2f55e3c2b5
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu May 26 11:54:57 2016 +0000

Function to create config-drive partition
    
    Add function to create partition for config-drive to util/disk.py
    module. This is required to switch from 'cobbler' to 'image'
    provision method.
    
    Change-Id: Ia52195a2e25e4e459c8947eb2dfaea97195dbe7f
    Related-bug: 1575054

commit 58b294c8c36da9398940c15a692283212da35861
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Mon May 30 19:43:02 2016 +0000

Verify network template sanity before loading
    
    Open file with network template and load it as YaML. Fail, log
    and raise exception if any error occurs opening the file.
    
    Change-Id: Ie516749c832c23bb739ebb5d02dd9bd5a9807b7f
    Closes-bug: 1583063

commit 86a7e912a3efe00dc686014d17060ab001638b8a
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Mon Jun 6 12:58:15 2016 +0000

Add function for checking sanity of nodes to upgrade
    
    Move sanity checks logic for upgrading node into separate
    function.
    
    Change-Id: Ifa46753f0bed18a0e25887c2a07445145011dad4
    Related-bug: 1583063

commit 4dccecc542c32e9120ae85c60bd74bdc04f18e85
Author: Yuriy Taraday <yorik.sar@gmail.com>
Date:   Tue Jun 21 17:59:47 2016 +0300

Remove helpers/docker_patch.py
    
    It is not used anywhere, so delete it.
    
    Change-Id: I26d999e419a8a9b066a995cd78dee9ab7910ef06

commit 9f1eca9bd6f2f87e606f3b657fcc34778ba34a83
Author: Ilya Kharin <akscram@gmail.com>
Date:   Tue Jun 21 02:10:21 2016 +0300

Stop ostf before to re-initialize master node
    
    The ostf service should be stopped before to apply all set of puppet
    tasks because the respective task does not work properly. To be sure
    that ostf will be started with an appropriate config it should be
    restarted.
    
    Related-Bug: #1594818
    Change-Id: I17d6fcd20909c0990b0acbeab1bd5bc3e9d2e98b

commit 6111d9b9de10b583c5a1699edcd0994e55a739e3
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu May 26 14:54:41 2016 +0000

Validate nodes roles to exclude upgrade of single controller
    
    The proper sequence to upgrade controllers is:
    
    - upgrade primary controller in isolation
    - cut over to the new control plane
    - upgrade remaining controllers at once
    
    To exclude the possibility to upgrade single controller after
    the cutover, add validation of roles of nodes to upgrade and
    stop if the only node passed to the script is a CIC.
    
    Change-Id: Ie91ae35acb7e36103004a59c92cb495837e0dca0
    Closes-bug: 1584656

commit a15d657e65bfde9211794d96df6d7c3fb2baa4ef
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Thu Jun 9 13:06:00 2016 +0300

Add helper functions for work with temp dirs and files
    
    * Add function get_tempname required for create temporary filename
    It has optional arguments: dir and prefix:
    dir setup a directory where this file will be created
    prefix setup a prefix value of the file
    
    Usage:
    
        from octane.util import tempfile
    
        tmp_name = tempfile.get_tempname()
    
    * Adding of contextmanager temp_dir is required for creating of tempdir
    before code block and removing it upon completion of the block.
    
    Usage:
    
        from octane.util import tempfile
    
        with tempfile.temp_dir() as my_tmp_dir:
            # code block
            ...
    
    Change-Id: I20b8a0109a85ef41708fa8c9e7e050b391ed039e

commit 40956ee786e08cac5ededb05f5bf3d2fb127236a
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Mon Apr 11 17:08:53 2016 +0300

Add not live migration case in upgrade node
    
    with argument --no-live-migration you may start command
    upgrade-node and upgrade more than 1 compute and more than 1 ceph-osd in
    one command
    
    Usage:
    
        octane upgrade-node --no-live-migration $SEED_ID <ID> <ID> <ID>
    
    Change-Id: I77fc7404131ebc8d30cf53a39c63e48c00481188

commit 18f2a6b09e2f06e558c2d7d6be53bf45d130797f
Author: Mykola Yakovliev <vegasq@gmail.com>
Date:   Tue Apr 26 15:14:18 2016 +0300

Delete old images after restoring
    
    Remove exists images after restore call because
    they are contains old ssh keys and can't be used
    anymore.
    
    Change-Id: Ib8887621d2fef69c2ee00ebad486415a9be476d3
    Closes-Bug: #1558057

commit 65e13c9a2a2edb015279f34d8965ee0c4a2d77d1
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu May 26 10:14:45 2016 +0000

Insert 'rgw_frontends' parameter with default value into ceph.conf
    
    In fuel-library, parameter 'rgw_frontends' was added to ceph.conf
    file in version 8.0. This parameter always takes default value
    and cannot be managed by deployment configuration engine (nailgun).
    
    Add a line containing this parameter to the ceph.conf file while
    moving it to upgraded controller. Set parameter to default value.
    
    Change-Id: I19866f6e9c1b7590c017f471cb11e78c72102136
    Closes-bug: 1575689

commit 44ba73e109e851e73937481e41640705a323f41c
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Jun 17 18:02:16 2016 +0300

Apply the keystone task with appropriate creds
    
    Change-Id: I8bbf1ed32136f77177f1f1be785138d02427b3fd

commit a3c76579516791e94e98cb17600cfadb1d56a9b3
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Fri Jun 10 19:48:15 2016 +0300

Refactoring in ceph osd handler
    
    This is required for mutiple upgrade ceph nodes
    fix problems:
    * setup patch for every node (now this is done only once)
    * set/unset noout for every node (now this is done for each env)
    
    Change-Id: Ie17545d9fa4d4f6083130af1219ccb3efda067fc

commit a5823254ddb853730669859e9802274099087c81
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Jun 15 15:57:12 2016 +0300

Remove containers related code from mirrors
    
    This patch runs SQL queries on the host side instead of doing this
    inside of the postgres container. This is needed as a part of the work
    for removing docker containers from the master node.
    
    Closes-Bug: #1592802
    Change-Id: Ied8dde10908b275399080308af3f00f440ba36a2

commit ccbe6604baac0f61e54c2feb794336d7d35581cb
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Apr 20 11:08:08 2016 +0300

Add re-initialization of the master node
    
    PuppetApplyHost is renamed to PuppetApplyTasks to perform full
    re-initialization of the master node by applying all puppet tasks.
    
    Change-Id: I52a02fd4e72c1fd6367f91f088beacc0e9715969

commit 3ca668f9df9454d59483588f09c889e87c933ef4
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Jun 10 12:28:27 2016 +0300

Remove docker related code from LogsArchivator
    
    Change-Id: Ife1f84c9e6a87649d143493103516929042067e5

commit a0075c500ddd648cca152e5f306086ea9a14f532
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Jun 10 02:49:49 2016 +0300

Create releases only if they are absent
    
    To support the backup/restore feature of a master node within one
    release already presented release should not be loaded twice. This
    change makes the ReleaseArchivator handler idempotent.
    
    Change-Id: Id5b3130fe7e4e836cac1661639052db25a48050f

commit ed32a78bc8cb0d75ac1c00d3ae00d2191e04c289
Author: Ilya Kharin <akscram@gmail.com>
Date:   Thu Jun 16 00:51:36 2016 +0300

Properly extend fixtures from openstack.yaml
    
    Fixtures should be properly extended from the 'extend' attribute for all
    embedded objects to support at least two nested levels. That is needed
    to cover two Ubuntu releases dependent on one another in 9.0 or some
    custom fixtures with a variable number of releases and their
    dependencies.
    
    Closes-Bug: #1592741
    Change-Id: Ib74cc0656e0f7fec3e1b3de3ac4cc4fd83e45d99

commit ba3574435f327a81a270f677f850acd2f68f504d
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Jun 10 02:42:16 2016 +0300

Remove docker related code from ReleaseArchivator
    
    Change-Id: I51f5ba59bb9b340152d832899f6290290d4e8b74

commit 1b4a4a771519ee127c215a949e080757710addab
Author: Ilya Kharin <akscram@gmail.com>
Date:   Thu Jun 16 21:42:50 2016 +0300

Remove code that sets the deployed_before flag
    
    The deployed_before flag is not more required attribute of cluster
    since the 9.0 release [1]. This patch removes the respective function
    that is responsible for setting the flag for all cluster as a part of
    the database restore procedure.
    
    [1] https://review.openstack.org/292804
    
    Change-Id: I22f4c5d9365af44326363de6bb57e132e1384c7a

commit cc6f6da8b8bd645bb84b3e2cdd00b1afefb55bde
Author: Ilya Kharin <akscram@gmail.com>
Date:   Thu Jun 9 19:12:41 2016 +0300

Remove containers specifics from DB archivators
    
    Since nailgun and keystone services are not isolated in containers their
    processes should be stopped before to manipulate with databases and
    data.
    
    DB archivators are collected in DatabasesArchivator to apply the
    'postgresql' task before to restore DB states. This task is needed to
    configure databases, users and privileges according new parameters in
    astute.yaml.
    
    Some new functions were introduced to perform respective operations on
    the host level, such as:
      - archivate.archivate_cmd_output
      - sql.run_psql
    
    Change-Id: I67145808e8c453e371fe20575c5cc6a729db139a

commit 07aefb822d69d4446014675f8b17667be2911149
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Apr 29 11:00:33 2016 +0300

Remove containers specifics from CobblerArchivator
    
    Some manipulations with services were removed from this archivator
    because puppet tasks will be re-used to re-initialize the master node.
    
    PathFilterArchivator was added to support the simillar functionality
    that ContainerArchivator provides but perform all operations in the
    host filesystem.
    
    Change-Id: I46c7f9980f1cb8696a340e663101bec1711f4247

commit 9f99b29f8e868b68984951cd493d189391d9a753
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Apr 20 10:42:38 2016 +0300

Remove docker related code from AstuteArchivator
    
    Since the removing docker containers from the host re-initialization of
    services should be done by applying puppet manifests. The code that
    responsible for doing this with containers were removed from
    AstuteArchivator and now this archivator responsible only for restoring
    the astute.yaml file, this re-initialization will be added as a separate
    archivator.
    
    Change-Id: Ied3b5132e9dce1c37b259f74d76c93f56dbcdf67

commit c61d6b0beb14d825a770abf49ec2e76ead4fbb7f
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Jun 10 00:20:29 2016 +0300

Add function to run puppet tasks on a master node
    
    Since the 9.0 release on a master node all services were moved from
    docker containrs into the system namespace. All of these services are
    initialized by puppet manifests that are organized in tasks for each
    component, such as cobbler, nailgun, keystone, etc. The added function
    provides an ability to apply one task by its name.
    
    Change-Id: I301f8cacae87d5a64d3e99a2b351c00663b6db0a

commit 99dae3f5559aa5ed2322e7750a2abc46468f446f
Author: Ilya Kharin <akscram@gmail.com>
Date:   Tue Apr 19 10:48:42 2016 +0300

Disable unsupported archivators
    
    Docker containers where removed in the 9.0 release on the mater node
    that is why unsupported archivators were disabled in this changes to be
    able re-write them one-by-one.
    
    A list of archivators that use containers and should be modified:
     -  astute.AstuteArchivator
     -  cobbler.CobblerArchivator
     -  puppet.PuppetArchivator
     -  postgres.KeystoneArchivator
     -  postgres.NailgunArchivator
     -  release.ReleaseArchivator
     -  logs.LogsArchivator
     -  puppet.PuppetApplyHost
    
    Change-Id: I62f56070855ad4d8165c167bca6df0a97ad798c9

commit 3d603903a587dbc7f427bb9296bb077c94573cca
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Jun 17 03:24:13 2016 +0300

Use OS_{USERNAME,PASSWORD} env variables for auth
    
    The deprecated environments variables, such as KEYSTONE_USER and
    KEYSTONE_PASSWORD, are no longer supported by the fuel client and were
    changed on the actual ones, such as OS_USERNAME and OS_PASSWORD, in this
    patch.
    
    Change-Id: Id015b0426ca8ba93a947c68d9d8e1e207f1b91ad

commit 7f9d271db41f43292503239291accffe87459312
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Fri May 27 14:23:55 2016 +0000

Replace ssh.call with ssh.call_output for listing partitions
    
    Going from ssh.call to ssh.call_output simplifies code and testing.
    
    Change-Id: I5d115ccd1ae69f6b11434ca8f3ba84ef3b3aa049
    Related-bug: 1575054

commit a2f94378877fbeed33bbbc6d8d373ec809e570dd
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Tue Jun 14 17:11:37 2016 +0300

Patch puppet
    
    * Add ignore_db_dirs config to puppet mysql manifest
    
    Change-Id: I6a777b2226aec07c823f3099f3c3b758fa056d2c
    Closes-bug: 1592373
    Related-bug: 1484552

commit a1b933a84347a6877c8ee4123fe23446f85d9006
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed Jun 15 15:49:42 2016 +0300

Add applied_patch context manager
    
    Context manager setup patch before block and revert patch upon
    completion of the block.
    
    Usage:
    
        from octane.util import patch
    
        with patch.applied_patch(
                "patch_directory",
                "patch_file_1",
                "patch_file_2",
                ...,
                "patch_fiel_n"):
            # code block
            ...
    
    Change-Id: Id7bc86c6db4e55a0758257f7ecbb2538b1765123

commit c45e4be0a88addf06821f7ff246f941799e6b6b1
Author: Pavel Chechetin <pchechetin@mirantis.com>
Date:   Mon Jun 6 16:50:11 2016 +0300

Update fuel_agent patch for Fuel 8.0
    
    Partial-Bug: 1575054
    Change-Id: I48ba3a79c30ca8b21b880dab7cbb196ae3cb0712

commit a48834ab24951ccbdc210bddc06fe5cfeb23a3c3
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Jun 8 15:05:19 2016 +0300

Separate restoring of logs into LogsArchivator
    
    The part that is responsible for restoring of logging files of nodes
    was separated from NailgunArchivator into a separate archivator that was
    called LogsArchivator.
    
    Change-Id: Ic2684165935d90ad19953919e8ec01be20564b35

commit 74ef2b95fdc30d4eb124716dab1b98885baf0701
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Jun 8 03:23:37 2016 +0300

Separate restoring of release in ReleaseArchivator
    
    The part that is responsible for restoring of an appropriate release
    was separated from NailgunArchivator into ReleaseArchivator.
    
    Also _post_restore_actions was decomposed on two separate functions:
      - _repair_database_consistency
      - _create_links_on_remote_logs
    
    Change-Id: I4a91fe8db962f45f8b2050e4772d2aafaf2060ff

commit 816c83eb36d6b94eb4f0602a62bf45684e31fc78
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu May 26 10:28:22 2016 +0000

Add upgrade_levels section to nova.conf if it is missing
    
    When no [upgrade_levels] section present in nova.conf file,
    fuel-octane should add it and set the upgrade levels. Otherwise,
    live migration is not possible for rolling upgrade scenario.
    
    Change-Id: I60e5f88beab3c5da039885522b17355cedde7582
    Closes-bug: 1583596

commit 7e36c446d748985072504cd675296b1004a08cf8
Author: Ilya Kharin <akscram@gmail.com>
Date:   Mon Jun 6 19:38:49 2016 +0300

Revert "Specify upper constraints for cryptography===1.3.4"
    
    This reverts commit e468289cfd5270ac41a8bd0d1368a4fc94d2060e.
    
    Change-Id: I37696349878f26e47d30eafd7fc712b52fa4622f

commit 6d7329330a89069bcdd5f3e3cedcf3f6cc20aac3
Author: Ilya Kharin <akscram@gmail.com>
Date:   Mon Jun 6 19:54:05 2016 +0300

Add libffi-dev libssl-dev as binary dependencies
    
    Due to the problem with the new version of cryptography==1.4 we need to
    add libffi-dev and libssl-dev in a list of binary dependency to properly
    prepare CI environments and unblock gates.
    
    Change-Id: I521776d156f8d69f060185d089adb817139300d6

commit e468289cfd5270ac41a8bd0d1368a4fc94d2060e
Author: Ilya Kharin <akscram@gmail.com>
Date:   Mon Jun 6 13:59:07 2016 +0300

Specify upper constraints for cryptography===1.3.4
    
    Due to the fact that cryptography===1.4 does not support the old version
    of OpenSSL<=0.9.8 we need to refuse the upstream upper-constrains.txt or
    create our own.
    
    Change-Id: I7fb9956a81a45e2069494ad50709d36ade5521e8

commit 138ebacebc65de61c474a71449bcf32d0f161346
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Tue May 31 11:35:38 2016 +0000

Move network utils to util/network.py module
    
    Currently network.py utility module lives under helpers/ module.
    Move the module to util/ and update imports.
    
    Change-Id: Ia9adc2185f2b9d2296954856fb37a0b9608bd9c3
    Related-bug: 1576150

commit d0250fe95eae8a48c3920a1e413c0d1008a435c8
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed May 25 13:03:48 2016 +0300

If dockerctl don't stop container stop it using docker
    
    After maintenance update dockerctl has problems in stop docker
    container.
    
    Change-Id: Ie95a47816d62c2a8c684659e75434057d47fad2d
    Closes-bug: 1585151

commit 258d16b0f3dd9ebb8c4466108709911192507b69
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu May 26 11:05:04 2016 +0000

Move disk.py module to util module
    
    Current location for disk.py modules is inconsistent with the
    usage of utils for basic operations on disks, not only their
    metadata. Non-metadata operations must be implemented by util
    module.
    
    Change-Id: Ia53b831f4b4d5c31da8801a8561b4721fd35d511
    Related-bug: 1585938
    Related-bug: 1575054

commit 62f75c10445a3cd275c86dae8028e8f444ef1968
Author: Roman Sokolkov <rsokolkov@mirantis.com>
Date:   Thu May 26 09:36:27 2016 +0300

Apply network templates before preupgrade handler
    
      Controller's "preupgrade" handler replaces
      deployment_info if "isolated". After that
      Octane applies network template.
    
      But nailgun will NOT apply any new settings
      if replaced_deployment_info exists for node.
    
      Closes-bug: 1572240
    
    Change-Id: I06922a59a5f6976acb7c268460fb95fcde86aa54

commit f73e0c7a9942c2fe211a82dd19107e9fa351e0bd
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu May 19 16:54:02 2016 +0300

Add enable-release command
    
    This command enables to deploy an older release of OpenStack
    with the upgraded Fuel Master node. Release to enable is
    identified by the ID. Administrator's password is required
    to access the Fuel API.
    
    Example:
    
      octane enable-release --id 1 --admin-password admin
    
    Partial-bug: 1567535
    Change-Id: I625db20e570b177cb46143818ac65c5d78cda0e2

commit 597c8c617b43c5ad05b3767672da96fa79fc12b6
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Wed May 25 10:03:24 2016 +0000

Include patches recursively in RPM package
    
    Parametrize the RPM spec with python2 macros and add
    patches to  the build directory recursively  to make
    sure that all of them are included.
    
    Change-Id: I49488776047b32b0a939bbbe435c0c286eabedee
    Closes-bug: 1585528

commit 114d42bf09f167750844ffd3214aef16dac88714
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Mon May 23 12:39:20 2016 +0000

Add octane/bin directory to manifest file
    
    The directory octane/bin contains non-python scripts and
    binaries that octane uses to perform certain tasks locally
    and on remote nodes.
    
    Change-Id: I867f88de636a7fe666cdb537b70f18f05e332508
    Closes-bug: 1584638

commit 34ece032c93df372e781dba1fc18f516e73d04b9
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Thu May 19 13:21:10 2016 +0300

Close fd created by mkstemp
    
    In PuppetApplyHost archivator file descriptor returned by mkstemp is ignored
    and thus remains open after function finishes.
    
    Change-Id: Ief11db89680082bec10da161277fdf1b2ff436a3
    Closes-bug: 1572918

commit ff7a897aa40453d58cc7e4b397b3d3f65db2b1ce
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed May 11 17:46:47 2016 +0300

Unhealthy ceph don't break upgrade
    
    Bad health ceph status shouldn't stop node upgrade.
    The reason is fuel overestimates ceph default value for pool
    placement groups number.
    
    Change-Id: Iff301ffeedd26a6f532dbbd480a6395cd774181a
    Related-Bug: 1464656

commit 969c90144e468942d54bcc0ccc7fa820b7b4a0a1
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Tue May 17 21:49:22 2016 +0000

Use shutil.copy2 to preserve file stats in backup
    
    To preserve file ownership and permissions in backup and
    some other places, move from using shutil.copy to shutil.copy2.
    
    Change-Id: I8f00834ab598267ccc54f18cea70d6db415ba69d
    Closes-bug: 1572922

commit c6dc22cdeaaffcd0212e5e031cfabf59c6405bb5
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Tue Apr 19 17:07:36 2016 +0300

Add full repo backup
    
    Required for backup repositories in master node if there aren't
    deployed environment.
    
    Usage:
    
        octane fuel-repo-backup --to backup_path --full
    
    Change-Id: Ic9560edd3390f63ff015d68c0b5747351b19cf66

commit 50826027d47bc8666f67799d03809d8266f59062
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Tue May 17 12:39:49 2016 +0300

Move get mirrors in backup to separate function
    
    NaigunWWWBackup class have the function _get_mirrors now.
    This is required for geting backup dir paths.
    It can be easely overload in child class.
    
    Change-Id: I0d8d9867249f4dd98c084a7aea02d26d742835fc

commit d4e7cc3b4f63c9e8d414d33977c574e4ed41a14e
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Tue May 17 12:32:08 2016 +0300

Change sql call in backup mirrors
    
    Change sql call from docker run in postgres conatiner to util
    run_psql_in_container function
    
    Change-Id: If1fa514f9b7d67c5d78cdfcc54bdba919ecc8834

commit a40e4e4073b05e825daa11fccf1c2735dc1f6cb4
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed May 4 16:53:27 2016 +0300

Change environment settings on upgrade-env cmd
    
    set up provision image_data uri based on release version (before that
    them copied from original env)
    
    Change-Id: Ibb7400d2b98ca81fd2a58ca76b04a3e80d4e76a5
    Closes-bug: 1582231

commit b1c430528edc5e56046cc822fe91de54da9cf578
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Mon Apr 25 18:03:37 2016 +0300

Move sql container runner into util function
    
    This function is used to execute SQL-queries by using psql
    in the postgres container.
    
    Usage:
    
        from octane.util import sql
    
        sql_query = ...
        sql.run_psql_in_container(sql_query, container, db)
    
    Change-Id: I180a0179f223fa59bd0a99cd90b093a153b6ddfe

commit 306321cbc924759c2cb31e6b05506f0ff22dc6da
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Wed May 11 14:44:55 2016 +0000

Restore fixes for partitioning schema for versions before 7.0
    
    Partitions schema created by fuel-agent for image-based provisioning differs
    from the schema used for cobbler provisioning. Cobbler provisioning was
    supported in versions prior to 7.0.
    
    If version of the environment is lower than 7.0 and provisioning method is not
    image-based, fix the partitions schema at the nodes being upgraded.
    
    Change-Id: I27a5477186efe62e3e3dc6f96b16889da1a49250
    Closes-bug: 1574624

commit 762d70e61d4852566386fb34e944dcae79bd6aac
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Wed May 11 14:40:56 2016 +0000

Remove fix for custom field in release metadata
    
    Remove the script that patches custom field in the release metadata from
    the code tree.
    
    This script was requried for upgrade of customized version of 5.1.1
    release of Fuel. Since that version is no longer supported, the fix
    is also not used anymore.
    
    We still need to prepare for later patching of a node's metadata by
    moving other scripts and patches to the location where they are expected
    to be.
    
    Change-Id: Ifb2590f96449716db09219240ceda6738951f95b
    Related-bug: 1574624

commit f3518dbe3901d13200b026206b391bec79aa3fb7
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Fri Apr 22 18:42:51 2016 +0300

Fix shutoff_vms cmd
    
    Make function shutodd_vms runable
    
    * Add close quote on awk format
    * Run command on controller node
    * Fix argument and subcommand order in nova list command
    
    Change-Id: I767950b35b8b85a76833f38d5049de60079e68f6

commit 685241e4b7ea8b0b9150b505d435c497722680f5
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu May 5 13:13:51 2016 +0000

Stop update of volume metadata gracefully if 'os' volume is not on LVM
    
    To upgrade an OpenStack cluster node with preserved partitions and data
    from version 6.0 to version 8.0, changes in disk metadata required. These
    changes accommodate the metadata to change of provisioning method from
    'cobbler' to 'image'.
    
    To avoid failure if the metadata was already altered by the previous
    execution of upgrade-node command, the script update_node_partition_data.py
    should stop gracefully if no relevant partitions were found.
    
    Change-Id: I7a110423312d1cd4d5101e7ce048847092d8578f
    Closes-bug: 1578639

commit 8616d029eccb366ee4a36d51e65d099ea190c69c
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Mon Apr 11 13:16:32 2016 +0300

Change check on host_evacuation
    
    Check number of enabled compute nodes on host_evacuation before disable
    current compute node.
    
    Change-Id: I98f972cb028425d89ae21f77ade2f2fa2e3ee1a5
    Closes-bug: 1563253

commit b608842c241db2f23b8c01f6153459b27832222b
Author: Ilya Kharin <akscram@gmail.com>
Date:   Tue Apr 19 22:25:40 2016 +0300

Add a function to change password in astute.yaml
    
    This change creates a separate context manager for setting
    FUEL_ACCESS/password in astute.yaml.
    
    Implements: blueprint upgrade-fuel-admin-node
    Change-Id: I2699711c760650f9f5c69e6fb61c13b6190e61b4

commit d4b073bbf43e3fba99503f76b84086ab751d106c
Author: Roman Sokolkov <rsokolkov@mirantis.com>
Date:   Tue Apr 19 18:22:10 2016 +0300

Makes cache_service_tenant_id compatible with Juno
    
    This change returns compatibility with the keystone client to obtain ID
    of the services tenant for old releases, clients are used according to
    the next releases of Fuel:
    
        keystone: < 7.0
        openstack: >= 7.0
    
    Closes-Bug: 1572019
    Co-Authored-By: Ilya Kharin <akscram@gmail.com>
    Change-Id: I64877ea484cf148b0eb7501bf2676db5388f892f

commit 5cbf3f7c2e5a33705f2dcc5c6a6152cf12dd87c0
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Mon Apr 25 10:19:22 2016 +0000

Change provisioning method to 'image' for 8.0
    
    The only supported method to provision environment in versions past
    7.0 (including 8.0, 9.0 and so on) is 'image' for image-based
    provisioning.
    
    Update the value of the corresponding parameter to be compatible
    to releases in question.
    
    Closes-bug: 1573100
    Change-Id: I8397f2460539a503884fd813f484dd78df8e352f

commit 0165988b8f5421c3bf2f0d56d13ea2827f0f7672
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed Apr 20 13:57:11 2016 +0300

Add cobbler profiles and distros backup
    
    During the upgrade from the 6.0 release configuration files of profiles and
    distributions for Cobbler should be preserved because this release provides
    Ubuntu 12.04 that is not presented in 8.0.
    
    Closes-bug: 1571586
    Change-Id: Id9323cbd5f5b73451e5d97ffd4191155c5f5dd5f

commit d4dc8b4f4855b5b47d73dbab87f77950d55ddfc5
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed Apr 20 13:14:40 2016 +0300

Add option for backup_name in ContainerArchivator
    
    backup_name allow you to set up backup dir name for container
    archivator. This is required for backup few directories from same
    container.
    
    usage:
    
    class MyArchivator_1(base.ContainerArchivator):
    
        backup_directory = "/dir/name_1/"
        container = "container_name"
        backup_name = "container_backup_name"
    
    class MyArchivator_2(base.ContainerArchivator):
    
        backup_directory = "/dir/name_2/"
        container = "container_name"
        backup_name = "container_backup_name_2"
    
    Change-Id: I00b3b03053a2a848c318bac3379d03c3b94dd2e0

commit 3636aa37930988d0ef7704618c45324e503b9d08
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed Apr 20 12:47:17 2016 +0300

Add base CollectionArchivator class
    
    This class required for archivation and restore groups of archivators.
    You can set up extra_pre_restore_check, extra_backup and extra_restore
    action. This actions will runn after base_actions of classes setting up
    in archivator_classes.
    
    Usage:
    
    class SimpleCollectionArchivator(CollectionArchivator):
    
        archivator_classes = [Archivator1, Archivator2, Archivator3]
    
    Change-Id: Ie534ffd5c06abd48837f84dfb495410dbf5afa61

commit c74a47b5ff0c67b40ea4d2a5163e298cd05b01df
Author: Yuriy Taraday <yorik.sar@gmail.com>
Date:   Thu Apr 21 14:59:58 2016 +0300

Honor upper-constraints
    
    Since we don't have many special requirements anymore, we can honor
    upper-constraints and avoid some gate breakages.
    
    Cap on pytest is dropped because latest version works just fine on 2.7
    and we don't support 2.6 anymore (in master). It gets overwritten by
    upper-constraints anyhow.
    
    Change-Id: I173a4db6d14e22cbe11fc790e537b8f226db65f3

commit 865985bddd7bc1c9c731d79e5fd2fdf0e89f98a0
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Apr 8 22:28:37 2016 -0500

Fix keystone context for fuelclient==9.0.0
    
    In the 9.0 release fuelclient was changed and the previously developed
    approach does not work anymore. This change also includes modification
    of test for the 8.0.0 version of the client.
    
    Change-Id: Id1813b2365729c630a1aae0f6dd6a44e666d323e

commit f66bb89093bef781aa1318aab0afc310a2360b0f
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed Mar 30 12:47:07 2016 +0300

Add upgrade ceph osd command
    
    This command required for upgrade all osds on compute in one moment.
    If they are in wrong version of osd on computes it breaks osd cluster.
    
    usage:
    
    octane upgrade-osd --admin-password $password $ORIG_ID
    
    Change-Id: I0b8199ef2f4d39541afa39f5543d4863f28517f0
    Blueprint: upgrade-cluster-7-to-8

commit 95db28139cc9d5c38a05b58f31142e542ff5e23d
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Fri Apr 15 15:40:48 2016 +0300

Fix problem with pq dump restore
    
    Copy data to psql stdin using shutil.copyfileobj
    
    Change-Id: Ia77d45d01ee9c3d11049c003e7640c086a8da20f
    Closes-bug: 1570371

commit 9ab7a95f8cc0b7da6811ffd2c8f154560f63b9cf
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Mar 4 15:14:42 2016 -0600

Add ability to change roles during upgrade nodes
    
    This change allows to specify new roles in the new environment using
    the `--roles` option:
    
        $ octane upgrade-node --roles=<role1>,<role2> <env> <node1> <node2>
    
    Change-Id: Ie4f6863047ba3a9bb76d7b22ceeaa70927ca9fc1
    Depends-On: Iedb20a904e58f5b9a86eb47de8d8d317dc3cc61b

commit ad3fdd19e882372234fb3fa8f5d2bd6e7aac18bf
Author: Ilya Kharin <akscram@gmail.com>
Date:   Wed Apr 13 09:54:53 2016 -0500

Allow to upgrade nodes without re-provisioning
    
    It's possible to reassign nodes from the upgrade environment to the seed
    environment without re-provisioning.
    
        $ octane upgrade-node --no-provision <env> <node1> <node2>
    
    Change-Id: I1fb88c6c7e0581034d8668373ecaa3ace2d085af
    Depends-On: Iedb20a904e58f5b9a86eb47de8d8d317dc3cc61b

commit 373108b5ae19ab4a11d4a7b7b414019733e95e3b
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed Apr 6 12:30:35 2016 +0300

change get_service_tenant_id function
    
    Using openstack cli tool instead of keystone cli tool.
    
    Change-Id: I9e79a11fd0552a7d3205fcc02e05be01b33961ac

commit 0508a5a070c6490084f29bc43c9a140517fec396
Author: Yuriy Taraday <yorik.sar@gmail.com>
Date:   Tue Apr 12 12:43:06 2016 +0300

Add binary requirements file
    
    Currently OpenStack CI isntalls too much requirements even for simple
    jobs like PEP8 (zookeeperd anyone?). This file should fix that.
    
    Change-Id: I9ecd50a38e1efc430699c6d9ebd1f39b13864fd9

commit 04c37007e91d75f00199ef5784774bd6aad07259
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Tue Apr 5 15:36:07 2016 +0300

Change keystone authorization
    
    In fuel-9 change authorization method to keystone on controller node.
    
    Change-Id: I93935648164e2ad203256360373f8029eed9d464

commit 7ca840271284bb3102d4a59db62c29b3410344ce
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Fri Apr 8 15:40:55 2016 +0300

Move read astute conf to function
    
    move read config function to helper function
    
    Change-Id: I98179a410b212035ab61e9ed491c5527a59759dc

commit d395d6d92e7f0ea6ec41fb05873b0f64a3eb5556
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Thu Mar 24 15:21:44 2016 +0300

Patch ceph manifests
    
    This patch fix problem with on upgrade compute with creation new one not
    required osd.
    
    Change-Id: I3a244b2f7dbf335aa9b727f37a92fdeb4e79397f
    Blueprint: upgrade-cluster-7-to-8

commit fb061f4ef61ee0c8868b16a8779af252a34fd926
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Thu Mar 24 15:52:44 2016 +0300

Remove unsupported commands
    
    Delete commands that are not required for upgrade from version 7 to version 8
    and above.
    
    Change-Id: Ibd30603b60eb8ca3b68eefb70be39937c5f9d584

commit a4ed2778ab73d3e058bed9c621222774d45bfe32
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Mon Apr 11 10:34:03 2016 +0000

Generalize the patch application procedure
    
    Current depth for the patch command (-p3) is tied to historical
    format of patches to Puppet manifests which used to be based from
    /etc/puppet/modules directory.
    
    Change the depth to -p1 generalize patching approach for other
    paths in the host file system.
    
    Change-Id: I8538af2192aa29987404fbe130934122997a2625

commit 4aabc7340059a318a143029dea913e94d748ee54
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Mon Apr 4 10:28:37 2016 +0000

Extract puppet modules patching into utils
    
    Add function to patch the puppet modules on the Fuel Master node
    to util/puppet.py module. Refactor handlers to work with util
    module for patching of puppet manifests.
    
    Change-Id: Ie7b43493b119bd46c9ac107ebda37c12520ea7a2

commit c746f2208948b6f5eee02cacbb8916196df8ce26
Author: Ilya Kharin <akscram@gmail.com>
Date:   Fri Apr 8 17:11:02 2016 -0500

Remove the unused fuel2 clone-ips command
    
    The clone IPs feature was not merged in the cluster_upgrade extention
    since the 7.0 release. This patch removes this unsed command.
    
    Change-Id: I874722bac68269d2d7dc6daf2723dd9f5274d178

commit 31246fa85da2f18512d300f83a1b5b73f247a371
Author: Oleg Gelbukh <ogelbukh@mirantis.com>
Date:   Fri Apr 8 14:21:01 2016 +0000

Move puppet modules patch function to utils
    
    Move function to patch puppet modules from prepare command module to
    util.puppet module.
    
    Change-Id: I555464d289bbca42924b5f8bf5f915c21bd77adc
    Blueprint: create-standalone-script-to-backup-masternode

commit d377cf04bff3d01a4b809fc53280532aee1aa314
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Wed Mar 23 15:58:25 2016 +0300

Fix update controlle handler network settingsr
    
    fix update controller handler network settings:
    * do not remove predefined networks
    * set up segment id to net04
    * set up physnet as empty string to net04_ext
    * set up network_type as local to neto4_ext
    
    Change-Id: I888867fc8fc0b3037c47ea2c2c958f553ec379d4
    Blueprint: upgrade-cluster-7-to-8

commit 0a8967f00e94f75f7a7fbe43bd59853a4c626627
Author: Sergey Ryabin <sryabin@mirantis.com>
Date:   Wed Apr 6 16:51:59 2016 +0300

Remove paramiko version requirement
    
    Change-Id: Ie7a187f37b8bbc9caf9981eefc1a8b2460529e21

commit fe614b7d2501fa6bd6fe4426724d34c7ad49b806
Author: Sergey Abramov <sabramov@mirantis.com>
Date:   Tue Apr 5 18:42:30 2016 +0300

Fix empty sql list
    
    Change-Id: Iabca69931e7119eb1e129de9e7ff9d83d798fa8c
    Closes-Bug: #1566393

commit e824b554f0df86495c147fefa99d85bceec1ca38
Author: Artem Roma <aroma@mirantis.com>
Date:   Fri Mar 18 14:16:48 2016 +0200

Add copying of VIPs to node upgrade procedure
    
    According to [1] seed cluster must have assigned nodes in order to
    copying of VIPs to take an effect. Thus needed command call added to
    upgrade node procedure.
    
    [1]: https://bugs.launchpad.net/fuel/+bug/1549254
    
    Change-Id: I5ce2f50333e8993fde317d2853395e0d83260a51
    Depends-On: I33670e8f2561be6fe18cec75bfc7ecc056ae2f6b
    Partial-Bug: #1552744

tags:

added: in-stable-mitaka

Report a bug

This report contains Public information

Everyone can see this information.

You are

Subscribing...

Edit bug mail

Other bug subscribers

Related blueprints

Upgrade MOS cloud from version 6.x to 8.0

Remote bug watches

Bug watches keep track of this bug in other bug trackers.