Bug 1928285 - [LSO][OCS][arbiter] OCP Console shows no results while in fact underlying setup of LSO localvolumeset and it's storageclass is not yet finished, confusing users
Summary: [LSO][OCS][arbiter] OCP Console shows no results while in fact underlying set...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: Console Storage Plugin
Version: 4.7
Hardware: Unspecified
OS: Unspecified
unspecified
medium
Target Milestone: ---
: 4.10.0
Assignee: Pranshu Srivastava
QA Contact: Martin Bukatovic
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2021-02-12 20:49 UTC by Martin Bukatovic
Modified: 2022-03-10 16:03 UTC (History)
6 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Cause: No loading prompt was present while the PVs were being provisioned/capacity was 0 TiB. Consequence: A confusing scenario for the user as they won't know if all the resources are available or still in loading state. Fix: Add a loader for the same. Result: Loader provides a feedback to the user if the PVs are still being provisioned/capacity is being determined, and if it encounters an error while doing so.
Clone Of:
Environment:
Last Closed: 2022-03-10 16:02:37 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)
screenshot #1: Create Storage Cluster view, during "Storage and Nodes" step, with an option to enable Arbiter mode (228.45 KB, image/png)
2021-02-12 21:07 UTC, Martin Bukatovic
no flags Details
verification screenshot (118.99 KB, image/png)
2021-11-26 21:35 UTC, Martin Bukatovic
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Github openshift console pull 10348 0 None open Bug 1928285: Display loading message when PVs are being provisioned. 2021-10-28 10:00:17 UTC
Red Hat Product Errata RHSA-2022:0056 0 None None None 2022-03-10 16:03:19 UTC

Description Martin Bukatovic 2021-02-12 20:49:51 UTC
Description of problem
======================

During "Create Storage Cluster" workflow in "Internal - Attached Devices" mode
(aka LSO), OCP Console acts as if there were no storage devices or nodes to
use, reporting no results and disabling the options at first, but later after a
while, updates the view and shows results of the detection.

This is confusing, esp. when one tries to enable Arbiter stretch cluster.

The fact that detection is not yet ready or ongoing should be clearly indicated
until the results are accurate.

Version-Release number of selected component
============================================

OCP 4.7.0-0.nightly-2021-02-09-224509
LSO 4.7.0-202102110027.p0
OCS 4.7.0-260.ci 

How reproducible
================

3/3

Steps to Reproduce
==================

1. Install OCP cluster on vsphere platform, with worker nodes having local
   storage block devices attached (to be used later via LSO)
2. Install LSO and OCS operators
3. Via OCP Console, start "Create Storage Cluster" flow, and select "Internal -
   Attached Devices" mode.
4. Configure Local Volume Set during "Create Storage Class" step
5. Try to enable arbiter mode during "Storage and Nodes" step

Actual results
==============

During "Create Storage Class" step, at first no suitable nodes with local
devices are shown. Then after a while, nodes are reported.

During "Storage and Nodes" step, at first storage class

Expected results
================

When a detection is not yet finished so that OCP Console can't present it's
final results yet, it should indicate that the details are being loading or
detected.

Additional info
===============

This could be confusing for first time users, especially when trying to use
arbiter mode, because during about first 10 or 20 seconds of "Storage and
Nodes" view, "Enable arbiter" option can't be enabled. Since there is "Advance
Subscription" notice for this feature, and it takes significant time for the
results to be shown in the page, customers could be mislead into thinking
that they don't have proper subscription, while the "enable arbiter" option
is actually available.

Comment 1 Martin Bukatovic 2021-02-12 21:07:22 UTC
Created attachment 1756713 [details]
screenshot #1: Create Storage Cluster view, during "Storage and Nodes" step, with an option to enable Arbiter mode

See screenshot #1 capturing Create Storage Cluster workflow during "Storage and Nodes" step, with an option to enable Arbiter mode. When one arrives there for the first time, arbiter mode can't be enabled no matter what. Customer not aware of this could reasonably suspect that it's because a problem with subscription or minimum requirements, as both details are covered in info boxed next to the enable arbiter option. This could send first time users to check details which may not be necessary to check again. Only those who will retry to click on enable arbiter option later again will notice that the option is not in fact blocked.

Comment 3 Nishanth Thomas 2021-02-13 08:17:28 UTC
Moving out of 4.7
Yes its takes time for leading the nodes. I remember there was a similar bz reported, will check and update here

Comment 4 Nishanth Thomas 2021-02-13 09:21:57 UTC
@mbukatov , sounds like a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=1901720 ??

Comment 5 Martin Bukatovic 2021-02-17 19:16:59 UTC
Looks like root cause of BZ 1901720 is mostly to blame here as well.

That said, this bug describes additional details:

- it describes use case with arbiter stretch cluster, where the problem could cause specific additional confusion
- the expected behavior is bit different, I expect that UI will be able to tell user that LSO plumbing work is not yet finished via some info box, so that one can distinguish empty results from ongoing detection

So for the tracking purposes, I would like to keep it open for qe team to make sure this use case is verified.

If you plan to fix the issue at once, you can set the older bz 1901720 as blocking of this bug.

Comment 6 Elad 2021-02-18 08:35:38 UTC
I don't see how bug 1901720 is related here.


Bug 1901720 is about a cosmetic issue in which it takes some time to find the disks for storagecluster creation

while here (bug 1928285), it is about the fact that you can create the storagecluster with a single deviceset only and for using more disks, the admin needs to perform add capacity operations

Comment 7 Elad 2021-02-18 08:48:36 UTC
sorry, please disregard my previous comment, I thought this bug is bug 1928319.

About the 1901720 and this one (1928285), IMO, if the root cause is the same, maintaining 2 BZs is redundant.

Comment 8 Martin Bukatovic 2021-02-22 15:48:20 UTC
Additional information:

> This could be confusing for first time users, especially when trying to use
> arbiter mode, because during about first 10 or 20 seconds of "Storage and
> Nodes" view, "Enable arbiter" option can't be enabled.

I retried this process and measured time it takes for arbiter mode to be
available when one arrives to "Storage and Node" view for the first time,
and the result is 65 seconds.

Comment 9 Nishanth Thomas 2021-02-24 10:21:45 UTC

*** This bug has been marked as a duplicate of bug 1901720 ***

Comment 10 Martin Bukatovic 2021-05-17 17:31:35 UTC
With OCP 4.8.0-0.nightly-2021-05-13-002125, I see that BZ 1901720 is addressed, but this one is not.

Based on this, I'm reopening the bug, because it can't be a duplicate of BZ 1901720 when it's fix doesn't have any effect here.

Comment 11 Martin Bukatovic 2021-05-17 17:41:21 UTC
That said, I see that BZ 1901720 is still in POST with it's patch not merged yet, so it's possible that I considered BZ 1901720 addressed because of other change, and that the actuall fix of BZ  1901720 will fix this bug as well.

Comment 12 Nishanth Thomas 2021-06-09 10:01:19 UTC
Yes, so I am marking this as a duplicate of BZ 1901720 again. Please verify it once original BZ is moved to ON_QA

*** This bug has been marked as a duplicate of bug 1901720 ***

Comment 13 Martin Bukatovic 2021-09-17 16:37:57 UTC
I disagree, moreover the issue is still not addressed.

What we need to make sure is that:

- let user know that localvolumeset is being created (something is going on)
- localvolumeset and it's storage class has been created (something has just finished)

Comment 14 Martin Bukatovic 2021-09-17 16:38:49 UTC
I just checked that this problem is still applicable to OCP/ODF 4.9

Comment 15 Martin Bukatovic 2021-09-17 16:50:17 UTC
One possible solution here would be to include additional step in the wizard, which would track the progress of setup for localvolumeset and it's storageclass.

Comment 22 Martin Bukatovic 2021-11-26 21:33:56 UTC
Testing with:

- OCP 4.10.0-0.nightly-2021-11-26-060537
- LSO 4.9.0-202111151318
- ODF 4.9.0-249.ci

And I see that when LVS starts to be provisioned before 3rd step "Capacity and nodes" is shown, Console reports that "PVC are being provisioned", in a similar way how it's done during disk discovery phase earlier.

>>> VERIFIED

Comment 23 Martin Bukatovic 2021-11-26 21:35:03 UTC
Created attachment 1843778 [details]
verification screenshot

Comment 26 errata-xmlrpc 2022-03-10 16:02:37 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Moderate: OpenShift Container Platform 4.10.3 security update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2022:0056


Note You need to log in before you can comment on or make changes to this bug.