Pass optional timeout to service commands

[Petr Malik]

Pass optional timeout to service commands

Service commands may take some time to return.
The current code does not pass any timeout value into the execute
function resulting in the default 30s value being used.
That value turns out to be too short for certain datastores
(Cassandra) on slower/heavily-loaded systems.

Trove already has a timeout (default=3min) for instance state changes that
should be used instead (this timeout has already been used when
polling for database states).

The service command execute function have been updated to accept an
optional timeout kwargs (like the shell command execute function).
The state change timeout is now being passed in from DB
start/stop/restart calls.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/291941
Committed: https://git.openstack.org/cgit/openstack/trove/commit/?id=87658dc4cc59f7c87d5631fbd6132b98384d396c
Submitter: Jenkins
Branch: master

commit 87658dc4cc59f7c87d5631fbd6132b98384d396c
Author: Petr Malik <email address hidden>
Date: Thu Mar 10 22:41:59 2016 -0500

    Pass optional timeout to service commands

    Service commands may take some time to return.
    The current code does not pass any timeout value into the execute
    function resulting in the default 30s value being used.
    That value turns out to be too short for certain datastores
    (Cassandra) on slower/heavily-loaded systems.

    Trove already has a timeout (default=3min) for instance state changes that
    should be used instead (this timeout has already been used when
    polling for database states).

    The service command execute function have been updated to accept an
    optional timeout kwargs (like the shell command execute function).
    The state change timeout is now being passed in from DB
    start/stop/restart calls.

    Change-Id: I1ef45b95d9238a52a26f630961657755216786cb
    Closes-Bug: 1556314

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/mitaka
Review: https://review.openstack.org/313598

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/313598
Committed: https://git.openstack.org/cgit/openstack/trove/commit/?id=547816c0fe6b2671378dd71744a614f20048b6ae
Submitter: Jenkins
Branch: stable/mitaka

commit 547816c0fe6b2671378dd71744a614f20048b6ae
Author: Petr Malik <email address hidden>
Date: Thu Mar 10 22:41:59 2016 -0500

    Pass optional timeout to service commands

    Service commands may take some time to return.
    The current code does not pass any timeout value into the execute
    function resulting in the default 30s value being used.
    That value turns out to be too short for certain datastores
    (Cassandra) on slower/heavily-loaded systems.

    Trove already has a timeout (default=3min) for instance state changes that
    should be used instead (this timeout has already been used when
    polling for database states).

    The service command execute function have been updated to accept an
    optional timeout kwargs (like the shell command execute function).
    The state change timeout is now being passed in from DB
    start/stop/restart calls.

    Change-Id: I1ef45b95d9238a52a26f630961657755216786cb
    Closes-Bug: 1556314
    (cherry picked from commit 87658dc4cc59f7c87d5631fbd6132b98384d396c)

[Doug Hellmann (doug-hellmann)]

This issue was fixed in the openstack/trove 5.0.1 release.

[Doug Hellmann (doug-hellmann)]

This issue was fixed in the openstack/trove 6.0.0.0b2 development milestone.