Alternative syncSecret approach #1030 #1032
Conversation
k8s-ci-robot
added
the
cncf-cla: yes
|
This change is |
aledbf
changed the title
|
ping @jcmoraisjr |
|
Coverage increased (+0.1%) to 44.102% when pulling 3d9559b4bdc4430d54ba89e110dd421d175783a7 on aledbf:fix-ssl-sync into 37375ae on kubernetes:master. |
aledbf
force-pushed
the
fix-ssl-sync
branch
2 times, most recently
from
d77355f
to
d016ebb
Compare
|
Coverage increased (+0.1%) to 44.164% when pulling d016ebba40889862e5b2f3e42af54d0034a124e1 on aledbf:fix-ssl-sync into 3495bfb on kubernetes:master. |
|
Coverage increased (+0.1%) to 44.15% when pulling d016ebba40889862e5b2f3e42af54d0034a124e1 on aledbf:fix-ssl-sync into 3495bfb on kubernetes:master. |
|
It's good but some events like change the secret name from some annotations doesn't trigger the cfg updating. Couldn't find the time yet to investigate this and create better exploratory tests. |
|
@jcmoraisjr changed, now it takes 3 seconds to detect the update and trigger a reload. Please test again |
|
Updating the secret name from annotations are still ignored and updating a non-CA or cert secret is broken on backend_ssl/syncSecret with "no keypair or CA cert could be found". Anyway this is still broken before these changes, so it's better than before and I think it's fine to merge. In the next 2 days I'll find the time to evolve my own tests to these scenarios and propose new changes. I wasn't expecting so big and deep changes in a codebase releasing beta versions. It's hard to follow master and merge it to my codebase. I'd watch and follow every single PR if you'd say this will continue to be frequent, otherwise let me know when such changes are going to be merged so I can give a feedback before it's merged to master. |
aledbf
changed the title
This should be the last change to the codebase. |
|
Is this fix in your 0.171 build on quay? |
No description provided.