-
Notifications
You must be signed in to change notification settings - Fork 4.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Build image change trigger silently fails when update is forbidden on free-int #14725
Comments
Testing locally works fine. Could be timing related, but i don't see how. |
@smarterclayton maybe related to cache/informers? do we have any master logs collected? |
No, but it was recreatable. We'd have to bump logs high.
|
do you have a dump of both buildconfigs and the imagestreams after the trigger didn't fire? |
as long as the imagestream actually got updated, i'd expect the buildconfig to eventually see the change and fire (due to reconcilliation) even if it didn't catch the initial event. I don't remember what our reconcile/relist interval is now though (we raised it from 2 minutes w/ the switch to informers as i recall) |
I'll give you access
…On Mon, Jun 19, 2017 at 3:26 PM, Ben Parees ***@***.***> wrote:
as long as the imagestream actually got updated, i'd expect the
buildconfig to eventually see the change and fire (due to reconcilliation)
even if it didn't catch the initial event. I don't remember what our
reconcile/relist interval is now though (we raised it from 2 minutes w/ the
switch to informers as i recall)
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#14725 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ABG_pyMQUJUSqKvcgRB1gRo0E9ZN2qKgks5sFstygaJpZM4N9UIE>
.
|
|
xref to similar issue: https://bugzilla.redhat.com/show_bug.cgi?id=1462131 |
@smarterclayton @tnozicka found that after he restart the master the dc will trigger, so I suspect a cache miss issue? |
that would imply we have a relist/reconciliation bug too though (unless the interval is just so large we didn't get to a relist?) |
@mfojtik @smarterclayton I am suspecting EDIT: This was the issue with Build to DC |
Turns out this was permission problems:
The permission attached to my account was dropped by the 3.5->3.6 upgrade or something later than that. @jupierce can you link the issue we're using to track lost permissions (the bugz) It's likely this should have fired event - "couldn't trigger image because of permissions" or something |
I guess this should be an event.... |
@smarterclayton Instances of missing data described here: https://bugzilla.redhat.com/show_bug.cgi?id=1461340 |
So the exact bug is:
|
This means controllers can't trigger builds unless they have permissions to create all build types, which I guess I can just give to it, since technically it had it before. |
Policy fix up in #14792 |
@smarterclayton why didn't this also affect s2i builds? i assume the image change controller didn't have permission to create/update those either... |
all users have those by default on those clusters, and the controller inherited it from "system:authenitcated" |
v3.6.114 (online version 3.5.0.20)
Had a chained build (output of one build tagged into another as image source). Image pushed, no build started.
The text was updated successfully, but these errors were encountered: