[RFE] Netapp make netapp_snapmirror_quiesce_timeout end user option

[Maurice Escher]

Hi,

currently netapp_snapmirror_quiesce_timeout is an option of replica promote that can only be set by the operator running the manila service, i.e. the cloud administrator.

There are situations where a high timeout may be good: safeguarding planned failovers to prevent data loss.
But there are also situations where no time should be waited at all: disaster stroke and I want to do an unplanned failover fast.

So there is no: one fits all. Let's move the decision up to the end user client whether the timeout should be at a certain higher value or zero.

I can imagine this being implemented as a force option to the replica promote API, since this is nothing that could be expressed via share type. One and the same share could get into both described example situations at different times, so it is not a describing attribute of the share itself, but rather a thing to provide at the time of calling promote (or some cleverness to determine, whether it is a disaster situation or not. but how to implement such goes beyond my imagination currently)

Thanks in advance for your thoughts about this,
Maurice

[Maurice Escher (maurice-escher)]

More thoughts after brainstorming with kpdev: instead of a force-option the option per call can also be giving the value of the quiesce wait time to be used for that particular promote operation.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.opendev.org/c/openstack/manila/+/868486

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/c/openstack/manila/+/868486
Committed: https://opendev.org/openstack/manila/commit/5f0f14a7f8b99ac008be4604bcaeec2dbc149d06
Submitter: "Zuul (22348)"
Branch: master

commit 5f0f14a7f8b99ac008be4604bcaeec2dbc149d06
Author: Kiran Pawar <email address hidden>
Date: Fri Dec 23 12:41:37 2022 +0000

    Add quiesce_wait_time option to share replica promote API.

    Currently netapp_snapmirror_quiesce_timeout is an option of replica
    promote that can only be set by the operator. There are scenarios where
    high timeout value set by operator does not fit well e.g. disaster
    stroke and user want to do an unplanned failover fast. Added new
    option 'quiesce_wait_time' to share replica promote API which allows
    to use specified wait_time(in seconds) instead of config option.

    Closes-bug: #2000171
    Change-Id: Ib02063ee8b82f7374cd89f90e7f24a845c6c7cd7

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/python-manilaclient 4.3.0 release.

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/manila 16.0.0.0rc1 release candidate.