Bug 7916 - Load information not shown when subcluster configuration is missing
Summary: Load information not shown when subcluster configuration is missing
Status: NEW
Alias: None
Product: ThinLinc
Classification: Unclassified
Component: Other (show other bugs)
Version: trunk
Hardware: PC Unknown
: P2 Normal
Target Milestone: LowPrio
Assignee: Bugzilla mail exporter
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2022-05-03 16:03 CEST by William Sjöblom
Modified: 2023-08-30 16:16 CEST (History)
1 user (show)

See Also:
Acceptance Criteria:

Attachments
tlwebadm status section with all subcluster configuration commented out (14.39 KB, image/png)
2022-05-03 16:03 CEST, William Sjöblom 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Description William Sjöblom cendio 2022-05-03 16:03:03 CEST 
Created attachment 1035 [details]
tlwebadm status section with all subcluster configuration commented out

Given non-existent subcluster configuration or one that does not declare any agents, you can start a session just fine, but tlwebadm and tlctl does not show any load information. A possible scenario where this would happen is if `agents=...' is mistyped as `agent=...'.

We fall back on an agent on 127.0.0.1 which is present in the load information sent to tlwebadm and tlctl but not presented. Error messages or warning is given by neither the services nor tlctl.

This has been observed in both 4.14.0 and 4.14.0post.
Comment 1 Samuel Mannehed cendio 2023-08-30 14:18:15 CEST 
Adjusting the title to avoid confusion.
Comment 2 Samuel Mannehed cendio 2023-08-30 14:20:17 CEST 
This bug applies to both Web Admin and tlctl.

In the new version of Web Admin you get an empty page saying:

> System Load
> The ThinLinc cluster consists of 0 agents in 0 subclusters.

And tlctl will won't print anything:

> $ sudo tlctl load list
> $

The problem is that we have a mismatch between the load balancer logic and the presentation of the load information. The presentation of the load information depends on a working subcluster configuration, while the load balancer is capable of falling back on 127.0.0.1 as the agent.
Comment 3 Samuel Mannehed cendio 2023-08-30 14:50:52 CEST 
(In reply to Samuel Mannehed from comment #2)
> The problem is that we have a mismatch between the load balancer logic and
> the presentation of the load information. The presentation of the load
> information depends on a working subcluster configuration, while the load
> balancer is capable of falling back on 127.0.0.1 as the agent.

This was slightly misleading. Web Admin and tlctl do not read the configuration directly, this was fixed in bug 7857. They correctly ask the load balancer for information. However, the LoadInfo object does not contain the fallback value for the agent in case the config is missing.
Note You need to log in before you can comment on or make changes to this bug.