http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c1 Dominique Leuenberger changed: What |Removed |Added ---------------------------------------------------------------------------- Assignee|bnc-team-screening@forge.pr |kkaempf@suse.com |ovo.novell.com | --- Comment #1 from Dominique Leuenberger --- Assigning to Klaus, who submitted this change into TW with sr#542192 -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 Martin Pluskal changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mpluskal@suse.com -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 Dominique Leuenberger changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dleuenberger@suse.com, | |Mihai.Dinca@suse.com -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c3 Dominique Leuenberger changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |okurz@suse.com Flags|needinfo?(dimstar@opensuse. | |org) | --- Comment #3 from Dominique Leuenberger --- https://openqa.opensuse.org/tests/540493/modules/salt/steps/1/src is the entire test code - very well possible that there is something wrong, but who better than you as the SALT Master would be able to tell? Keep in mind that this worked until the day the new salt package landed; up to that date, the test worked as-is. CCing Oliver Kurz, maintainer of the openQA test (according the test metadata) -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c5 Michal Nowak changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mnowak@suse.com --- Comment #5 from Michal Nowak --- Maybe we see two different bugs? On s390x (https://openqa.suse.de/tests/1314594#step/salt/20) and x86-64 (https://openqa.suse.de/tests/1315003#step/salt/20) there was a minion key which got accepted by the master but on aarch64 (https://openqa.suse.de/tests/1314725#step/salt/20) and Hyper-V (https://openqa.suse.de/tests/1313339#step/salt/20) there wan't a key from minion. With Hyper-V we tend to have network issues/misconfiguration which prevent us to run some network related tests, aarch64 might be the same. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c7 Joaquín Rivera changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jeriveramoya@suse.com --- Comment #7 from Joaquín Rivera --- Once we start the minion service, sometime there is not enough time for the minion's public key to get stored in the master and during the pinging process sometimes is not able to ping the minion so quickly. I have a pull request adding two delays, one for each failure that we see, BUT it would be really justified to add these delays server-client communication when salt is used in automated-fashion like in our script, where we simply install packages, start services, accept keys and ping nodes? is a problem of the script that is not taking into account this latency and now is visible because of the python3 or we are overlooking something else here? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c9 --- Comment #9 from Joaquín Rivera --- Increasing to 30s works for the first problem, listing the key, but for the pinging progress still failing in some scenarios using: salt '*' test.ping -t 30 It looks too much time in my opinion in the connections (taking into account that master and client are in the same machine). Any suggestion? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c25 Oliver Kurz changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(okurz@suse.com) | --- Comment #25 from Oliver Kurz --- Implemented a workaround with https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/4317 which retries to ping the minions in case the first try failed. See it in action in https://openqa.suse.de/tests/1444576#step/salt/22 . This makes it not a test blocker anymore - the test should have had it - but does not resolve the original problem. That persists. Currently I do not know how I can efficiently help in this regard so I do not plan to continue working on this bug for now. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c33 Oliver Kurz changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|INVALID |--- Assignee|Mihai.Dinca@suse.com |okurz@suse.com --- Comment #33 from Oliver Kurz --- Well, that might be true if you just take a look at salt but looking at the complete system there *is* a clear regression: What worked in before does not work anymore. I'm not saying it's salt but we have to think from the system perspective. Regarding the bug is INVALID is maybe giving the wrong impression that DimStar and me were wrong in the first place. IMHO WORKSFORME is a better resolution. I will take this bug to ensure that the tests just accept the current behaviour as given and not produces soft-failures. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c35 Mihai Dincă changed: What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo?(okurz@suse.com) --- Comment #35 from Mihai Dincă --- By looking at the first comment the bug was implying salt is the cause

The latest modification for salt was the 'reset of the package' and the move to python3

No proof was provided in this regard. After spending a lot of time to prove that the assumption is wrong. The conclusion was that the bug is INVALID, meaning that: salt is not the cause of this. Please feel free to investigate and fix the actual problem. I can see that you are still not waiting for the key to be accepted and you do the ping right away: https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/4380 I have instructed you to do so here: https://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c11 and here: https://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c30 Why you are not doing it? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1069711 http://bugzilla.opensuse.org/show_bug.cgi?id=1069711#c37 Oliver Kurz changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution|--- |FIXED Flags|needinfo?(okurz@suse.com) | --- Comment #37 from Oliver Kurz --- geez, you don't have to fight me. I did not say that salt is at fault here. The bug subject says "Salt test fails since switch to python3" and this is true. The salt *test* fails, not *salt*. ok, setting to RESOLVED now. Yes, of course the test could be adapted but I don't see a need to touch it more right now. We can revisit it later. -- You are receiving this mail because: You are on the CC list for the bug.