Bug 1590391 - Describe refresh intervals, default values and visible consequences for the user
Summary: Describe refresh intervals, default values and visible consequences for the user
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: doc-RHGS_Web_Administration
Version: rhgs-3.4
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
: RHGS 3.4.0
Assignee: Rakesh
QA Contact: Martin Bukatovic
URL:
Whiteboard:
Depends On:
Blocks: 1503142 1590317
TreeView+ depends on / blocked
 
Reported: 2018-06-12 14:31 UTC by Martin Bukatovic
Modified: 2018-09-10 15:44 UTC (History)
9 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2018-09-10 15:44:30 UTC
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Bugzilla 1589321 0 unspecified CLOSED Cluster not available after unmanage 2021-02-22 00:41:40 UTC
Red Hat Bugzilla 1595360 0 unspecified CLOSED WA UI does not reflect the node down status when I reboot a node in the cluster 2021-02-22 00:41:40 UTC

Internal Links: 1589321 1595360

Description Martin Bukatovic 2018-06-12 14:31:47 UTC
Description of problem
======================

BZ 1589321 states:

> When cluster is unmanaged then it takes several seconds before it is listed in
> get cluster API call and available to import again.

During discussion ("RHGS-3.4.0 In-flight bug triage" meeting on 2018-06-12)
about this problem, it turned out we need to document what are refresh intervals,
it's default values and few examples of user visible effects (one mentioned in
the BZ mentioned above would be good enough).

Neha in https://bugzilla.redhat.com/show_bug.cgi?id=1589321#c1 provided some
details:

> UI is polling the cluster list API call after every 10 seconds. Hence, the
> updated data will be reflected in UI by maximum 10 seconds if the backend/API
> already contains the updated data.

Comment 1 Martin Bukatovic 2018-06-12 14:43:34 UTC
Some technical details are missing, and will be provided as an answer to
https://bugzilla.redhat.com/show_bug.cgi?id=1589321#c3

Comment 7 Martin Bukatovic 2018-08-07 09:23:24 UTC
Acking based on detailed description provided by Shubhendu in comment 4.

Comment 11 Martin Bukatovic 2018-08-17 15:15:47 UTC
BobbGt asked me for details, but I'm unable to provide suggestions outright.

I need to think this over and validate my understanding with Shubhendu or
Nishanth before going on. We need to be clear about impact of these intervals,
but to fix it, we need to identify the gap and the right answer *exactly* first.

Comment 12 Martin Bukatovic 2018-08-20 15:36:43 UTC
See also BZ 1595360 (it slipped through my attention because it's already
closed), where Nishanth states:

> Reboot will not reflect on the UI as the node comes back quickly. TTL for
> nodes is 150 seconds and the status will be reflected in the next sync. So if
> you shutdown the node, you will see that reflected in UI as expected after the
> TTL expires.

Comment 13 Nishanth Thomas 2018-08-27 05:25:09 UTC
For volumes it is sync_interval(Default is 60) + 450

Comment 20 Martin Bukatovic 2018-08-31 15:33:21 UTC
Since we have good eng. details for 2 use cases which has important visible
effects for storage administrator, I'm going to verify that these 2 cases has
been documented:

 * import of cluster again, just after unmanage
 * effects of TTL for volumes when Gluster storage nodes are shut down
   or offline

We don't have good end user description for other eng. details related to
sync intervals, nor were other use cases with important consequences for
the user identified during work on this BZ.

Both scenarios has been covered in Monitoring Guide:

 * import after unmanage is described in note in 3.1. "Unmanaging Cluster"
   section (as is noted in comment 8)

 * how TTL for volumes affects WA and dashboard when storage nodes are
   offline is described in "8.3. Volume Level Dashboard" section (my previous
   points about missing unit and value of sync interval has been addressed)


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