Project

General

Profile

Actions

action #175051

open

coordination #169654: [epic] Create test scenarios for Agama

Enhance test_running_installation.ts to finish earlier than 15 mins when failing

Added by JERiveraMoya 4 months ago. Updated 24 days ago.

Status:
Workable
Priority:
Normal
Assignee:
-
Target version:
-
Start date:
2025-01-07
Due date:
% Done:

0%

Estimated time:

Description

Observation

Avoid waiting 40 mins when failing, for instance:
openQA test in scenario sle-agama-10.0-agama-installer-ppc64le-sles4sap_default_unattended@ppc64le-hmc fails in
agama

Acceptance criteria

  • AC1: Provide mechanism to check in intervals if we are still on track installing to abort early when not.

Additional information

We can inspire ourselves with what was done in Playwright:
https://github.com/jknphy/e2e-agama-playwright/blob/main/pages/installation-finished-page.ts#L12-L23

Actions #1

Updated by rainerkoenig 4 months ago

  • Status changed from Workable to In Progress
  • Assignee set to rainerkoenig
Actions #2

Updated by rainerkoenig 4 months ago

Restarted the failed job to see what is going on, but the retry passed the agama test, but then failed later in validate_product.

Actions #3

Updated by JERiveraMoya 4 months ago · Edited

rainerkoenig wrote in #note-2:

Restarted the failed job to see what is going on, but the retry passed the agama test, but then failed later in validate_product.

Cannot reconnect, because the Agama profile we import doesn't have the post script to allow loging root, https://openqa.suse.de/tests/16380889/settings/yam/agama/auto/sap_default_no_scc.json
that is why is pointing to Lemon's ticket, but feel free to add it. Although for your issue I think you can reproduce it somehow because the repos are not reachable from ibs frequently.

Actions #4

Updated by JERiveraMoya 4 months ago

As a high priority ticket, please let's try to report feedback more frequently.

Actions #5

Updated by JERiveraMoya 3 months ago

I can see this WIP PR after one week of work which is a copy/paste of the example I provided in the ticket description.
Could you please share if you find some difficulties with this ticket that we could discuss aside (retro or 1:1) ? Happy to help if I know how...

Actions #6

Updated by JERiveraMoya 3 months ago

  • Status changed from In Progress to Workable
  • Assignee deleted (rainerkoenig)

Moving to workable and removing assignee to rotate in the squad as agreed in daily.
We need two things:

  • Being able to see what happen in the screen when is waiting so long, if we have some visual thingy that would help us to automate
  • Implement active wait that check every X minute that that thing is still there meaning that the installation in progress.
    Additionally consider the effort on this, as the monitoring tool for installation progress will arrive in next months.
Actions #7

Updated by jfernandez 3 months ago

I did a light research, maybe there is a way to manage that journalctl -u agama -f will show step by step the packages installed (In real time), maybe we can check this output in some manner.

Actions #8

Updated by JERiveraMoya 3 months ago

jfernandez wrote in #note-7:

I did a light research, maybe there is a way to manage that journalctl -u agama -f will show step by step the packages installed (In real time), maybe we can check this output in some manner.

all the 40 mins stuck situation I saw in failure were related with installation not starting, so we are not even try to look for weird things like disconnections, it is more like check that the screen that does the installation and install the package is there and if not exit early.

Actions #9

Updated by jfernandez 3 months ago · Edited

all the 40 mins stuck situation I saw in failure were related with installation not starting, so we are not even try to look for weird things like disconnections, it is more like check that the screen that does the installation and install the package is there and if not exit early.

In PPC64LE there is no screen to check, puppeteer is headless and there is no installation/displaymanager screen to check Agama status.

You could do a pre-check previous to start install but it will fail in repository error or package installation error.

Actions #10

Updated by JERiveraMoya 3 months ago

jfernandez wrote in #note-9:

all the 40 mins stuck situation I saw in failure were related with installation not starting, so we are not even try to look for weird things like disconnections, it is more like check that the screen that does the installation and install the package is there and if not exit early.

In PPC64LE there is no screen to check, puppeteer is headless and there is no installation/displaymanager screen to check Agama status.

You could do a pre-check previous to start install but it will fail in repository error or package installation error.

the page is still there in headless mode even if you don't see it, but of course the code should be done in puppeteer not with needles and openqa api.

Actions #11

Updated by jfernandez 3 months ago

the page is still there in headless mode even if you don't see it, but of course the code should be done in puppeteer not with needles and openqa api.

As I said in this approach you don't cover installation issues, only a pre installation check, this check can be included in all architectures and scenarios as a good practice but won't solve the issue in part of the cases.

Actions #12

Updated by JERiveraMoya 3 months ago

  • Tags changed from qe-yam-jan-sprint-fy25 to qe-yam-feb-sprint-fy25
Actions #13

Updated by JERiveraMoya 3 months ago · Edited

Mitigation: https://github.com/jknphy/agama-integration-test-webpack/pull/76 to 15 minutes, but still can exit early if we find a way with this ticket.

Actions #14

Updated by JERiveraMoya about 2 months ago

  • Tags changed from qe-yam-feb-sprint-fy25 to qe-yam-mar-sprint-fy25
Actions #15

Updated by JERiveraMoya about 2 months ago

  • Subject changed from Enhance test_running_installation.ts to avoid waiting 40 mins when failing to Enhance test_running_installation.ts to finish earlier than 15 mins when failing
Actions #16

Updated by JERiveraMoya about 1 month ago

  • Priority changed from High to Normal
Actions #17

Updated by JERiveraMoya 24 days ago

  • Tags changed from qe-yam-mar-sprint-fy25 to qe-yam-apr-sprint-fy25
Actions

Also available in: Atom PDF