Bug 56562 - UMC error message during long time operations
UMC error message during long time operations
Status: CLOSED FIXED
Product: UCS
Classification: Unclassified
Component: UMC (Generic)
UCS 5.0
Other Linux
: P5 normal (vote)
: UCS 5.0-5-errata
Assigned To: Marius Meschter
Florian Best
https://git.knut.univention.de/univen...
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2023-09-09 08:37 CEST by Stefan Gohmann
Modified: 2023-09-20 17:56 CEST (History)
2 users (show)

See Also:
What kind of report is it?: Bug Report
What type of bug is this?: 5: Major Usability: Impairs usability in key scenarios
Who will be affected by this bug?: 4: Will affect most installed domains
How will those affected feel about the bug?: 4: A User would return the product
User Pain: 0.457
Enterprise Customer affected?:
School Customer affected?: Yes
ISV affected?:
Waiting Support:
Flags outvoted (downgraded) after PO Review:
Ticket number: 20230908510006
Bug group (optional):
Max CVSS v3 score:


Attachments
management-console-module-appcenter.log (1.52 MB, text/plain)
2023-09-09 08:37 CEST, Stefan Gohmann
Details
management-console-server.log (1.16 MB, text/plain)
2023-09-09 08:37 CEST, Stefan Gohmann
Details
appcenter.log (880.87 KB, text/plain)
2023-09-12 11:27 CEST, Stefan Gohmann
Details
ucr_dump.txt (43.61 KB, text/plain)
2023-09-12 11:28 CEST, Stefan Gohmann
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Stefan Gohmann univentionstaff 2023-09-09 08:37:12 CEST
Created attachment 11115 [details]
management-console-module-appcenter.log

There are several support tickets about UMC error messages during long time operations. I was able to reproduce it during the OX App Suite installation.

The german error message during the installation is:
> Ein Fehler trat während des Verbindungsaufbaus zum Server auf. Bitte versuchen 
> Sie es später noch einmal.

UCS version: UCS: 5.0-4 errata794

I've attached the log files with UMC debug level 4. The error message appeared a little bit before 08:27:50.
Comment 1 Stefan Gohmann univentionstaff 2023-09-09 08:37:42 CEST
Created attachment 11116 [details]
management-console-server.log
Comment 2 Florian Best univentionstaff 2023-09-11 12:19:14 CEST
We see in the server log:

> 09.09.23 08:27:45.722  MAIN        ( ALL     ) : reset_timeout(): new session expiration in 28799 seconds
> 09.09.23 08:27:45.723  RESOURCES   ( INFO    ) : Searching for module providing command appcenter/progress> 09.09.23 08:28:59.823  MAIN        ( INFO    ) : The module appcenter is inactive for too long. Sending shutdown request to module


And inside of the appcenter log:
> 09.09.23 08:28:59.823  MODULE      ( INFO    ) : Received SIGALARM
> 09.09.23 08:28:59.823  MODULE      ( INFO    ) : Received SIGALARM
> 09.09.23 08:28:59.824  MODULE      ( PROCESS ) : Committing suicide

Looks like the reset of the inactivity timer is broken?
From a first look the logic look good: https://github.com/univention/univention-corporate-server/blob/5.0-4/management/univention-management-console/src/univention/management/console/resources.py#L316-L345

@Marius: can you have a look?
Comment 4 Stefan Gohmann univentionstaff 2023-09-12 11:27:29 CEST
Created attachment 11118 [details]
appcenter.log
Comment 5 Stefan Gohmann univentionstaff 2023-09-12 11:28:43 CEST
Created attachment 11119 [details]
ucr_dump.txt
Comment 7 Marius Meschter univentionstaff 2023-09-19 14:55:58 CEST
During the installation of an app apache2 is sometimes restarted.
This leads to the appcenter/progress request being aborted and an error shown to the user. 
However the installation continues in the background without any issue.

With this commit now there is a way to retry progress commands when they
have been aborted (aka HTTP status 0). A progress request will be retried 6 times before an error is shown.

This problem also occurs on UCS 5.0-3, prior to the UMC server unification.

Since git:329db0c5b700b {en,dis}able-apache2-umc does not prevent Apache 2 from being restarted during app installations. The OX-Appsuite installation restarts apache in its joinscript, causing some progress bar requests to be aborted.

univention-web.yaml
3fbdb03be80d | fix(web): retry aborted progress requests

univention-web (4.0.5-2)
3fbdb03be80d | fix(web): retry aborted progress requests
Comment 8 Florian Best univentionstaff 2023-09-20 14:03:10 CEST
OK: AppSuite installation doesn't fail anymore
OK: frontend-only fix: we could not re-activate the behavior of {en,dis}able-apache2-umc because then the OX joinscript fails
OK: stopping apache for multiple seconds doesn't break an app installation
OK: retrying of requests is delayed exponential
OK: progressbar breaks up when the server is not reachable for 3 minutes
OK: YAML