Bug 1459926 - Zaqar websockets randomly disconnect
Summary: Zaqar websockets randomly disconnect
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: rhosp-director-ui
Version: 11.0 (Ocata)
Hardware: Unspecified
OS: Unspecified
high
high
Target Milestone: ---
: 12.0 (Pike)
Assignee: Dan Trainor
QA Contact: Arik Chernetsky
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2017-06-08 15:09 UTC by Udi Kalifon
Modified: 2019-05-16 15:55 UTC (History)
9 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2019-05-16 15:55:28 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Launchpad 1803422 0 None None None 2018-11-16 15:15:53 UTC

Description Udi Kalifon 2017-06-08 15:09:46 UTC
Description of problem:
While working with the GUI, every so often an alert pops up: "Zaqar Websocket encountered Error". The easiest way to recreate it is to log in and out many times (it most often occurs on logins - maybe because it's heavily used in that stage when all the data is loaded?). 

This error also occurs during operations that take long, like role assignment or nodes registration, and when that happens it can cause the GUI to never see that the operation already finished (the spinner on the screen keeps spinning until you hit F5).


Version-Release number of selected component (if applicable):
openstack-tripleo-ui-3.1.0-9.el7ost.noarch
openstack-zaqar-4.0.0-2.el7ost.noarch


How reproducible:
Randomly

Comment 3 Jason E. Rist 2017-10-12 16:27:41 UTC
Patches have all landed. Need another review of this bz.

Comment 4 Honza Pokorny 2017-10-16 11:43:59 UTC
Closing this as stale; I haven't seen this bug in months now, feel free to reopen if you encounter it again

Comment 5 Ian Main 2018-03-07 18:47:06 UTC
We just saw this again in OSP 12 director UI.  This occurred while waiting for introspection to complete.

openstack-tripleo-ui-7.4.3-4.el7ost.noarch
python-zaqarclient-1.7.0-1.el7ost.noarch
openstack-zaqar-5.0.0-3.el7ost.noarch
puppet-zaqar-11.3.0-3.el7ost.noarch

Comment 6 Ian Main 2018-03-07 18:58:58 UTC
Saw this again while tagging nodes.

Comment 9 Jason E. Rist 2018-03-09 17:20:40 UTC
I wonder if this is something we need to rectify using quickstart?

Comment 10 Honza Pokorny 2018-03-09 18:14:35 UTC
In general, I think the nature of the websockets is such that it will disconnect at some point.  When you close your laptop, when you switch wifi access points, when you connect an ethernet cable, etc.  It's not so random.  The network connection isn't eternal or indestructible.

I'll look into automatic reconnects, and notifying the user when the connection is broken.

Comment 12 Dan Trainor 2018-11-13 23:48:48 UTC
Working on a related bug[0], I am no longer seeing these random disconnects by adding 'disablereuse=on'[1] in the Proxy member of mod_proxy for the Zaqar server that Apache httpd proxies for.

I'm continuing to test this for performance and compatibility but so far, I am not able to reproduce random Zaqar failures even with a fair amount of abusing UI (forced random refreshes, multiple actions on the same page, etc).



[0] https://bugzilla.redhat.com/show_bug.cgi?id=1647038
[1] https://httpd.apache.org/docs/2.4/mod/mod_proxy.html

Comment 13 Beth White 2019-05-16 15:55:28 UTC
The GUI is no longer supported in OSP14. The last supported release for the GUI is OSP13. Closing as won't fix.


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