No Graphs after disaster-recovery

statistics

(Mark Verlinde) #1

NethServer Version: 7.5.1804
Module: Statistics

(Would put this in the category minor bug / quirk, bug fits best…)

During testing disaster-recoveries noticed a little quirk in the Statistics Module. After a fresh install and in the first config wizard uploading a backup-config-tar.xz for the recovery procedure the Graphs are not shown:

signal-event nethserver-cgp-update solves the issue

EDIT:

Deducted it a bit more: restarting httpd-admin does the Job too;
I’ll gues reload could be enough (will try to remember to test next time…)


(Giacomo Sanchietti) #2

That’s strange, the post-restore-config event fires all package events using system-adjust.
And the nethserver-httpd-admin-update event already contains the restart (/etc/e-smith/events/post-restore-config/S70system-adjust).

Maybe there is a small race condition somewhere …


(Mark Verlinde) #3

Can’t find a something in the logs, can the [INFO] service httpd-admin is masked: skipped be a clue?

... esmith::event[6473]: Action: /etc/e-smith/events/nethserver-cgp-update/S00initialize-default-databases SUCCESS [0.777553]
... esmith::event[6473]: expanding /etc/httpd/admin-conf.d/cgp.conf
... esmith::event[6473]: expanding /etc/httpd/conf.d/cgp.conf
... esmith::event[6473]: Action: /etc/e-smith/events/actions/generic_template_expand SUCCESS [0.301748]
... systemd: Reloading.
... esmith::event[6473]: [INFO] service httpd reload
... systemd: Reloaded The Apache HTTP Server.
... esmith::event[6473]: [INFO] service httpd-admin is masked: skipped
... esmith::event[6473]: Action: /etc/e-smith/events/actions/adjust-services SUCCESS [0.567659]
... esmith::event[6473]: Event: nethserver-cgp-update SUCCESS

(Giacomo Sanchietti) #4

Yes, it should be the reason: to avoid UI breaking when restoring the a backup from the Server Manager.

I don’t know if there is a workaround for it. :thinking:


(Rob Bosch) #5

:hushed:
Now THAT would be a problem…


(Mark Verlinde) #6

Lets document the human-workaround until a more elegant solution pops-up.

Restart httpd-admin in the UI: Services > httpd-admin (restart)

and refresh the web page in your browser, you may have to login in again.


(Mark Verlinde) #7

Some new insights:
It does not occur if you go through the First-config-wizard and upload the backup-config in the UI.
(As described here)

Which let me think a possible solution-direction can be to restart/reload httpd-admin (unmask it) at the end of the First-config-wizard (transition to the servermanger-UI) in case of a disaster-recovery procedure.