Jack Lin
ref: https://github.com/longhorn/longhorn-tests/blob/v1.5.x/docs/content/manual/pre-release/node-not-ready/node-disconnection/node-disconnection.md (Case 2-1) there is no replica on the attached node + no data writing during the disconnection -> other replicas are still healthy -> auto-reattach after the disconnection...
ref: https://ci.longhorn.io/job/private/job/longhorn-e2e-test/1810/parameters/ This is a successful `network_disconnect` test
## Analysis ### According to the test logs - The test started at `2024-10-16T09:06:15.859007956Z Node Disconnect And No Replica On Disconnected Node` - The volume was attached to ` ip-10-0-2-134`...
I suggest we can use different volume/pod/resource name in each test To reduce the chance of interference and also make the debug easy It is hard to look up the...
Hi @yangchiu Did we encounter this issue before?
PR: https://github.com/longhorn/longhorn-manager/pull/2824
Hi @derekbit The PR is ready for review here: https://github.com/longhorn/longhorn-manager/pull/2824
Oh I already put it in previous comment
we are not going to backport this feature to v1.6.x The limitation is already in the doc https://longhorn.io/docs/1.6.2/advanced-resources/system-backup-restore/backup-longhorn-system/ Warning: Longhorn does not backup BackingImages. We will improve this part in...
Hi @a110605 Just discuss with me for more details or if you have any question, thanks!