Derek Su

With a cron job, we can calculate the checksums and compare the snapshots periodically for an active volume. The volume-head, however, might not be able to do it since the...

For the **silent corruption** caused by the bit rot, I don't find a good emulation method. I will update the test steps if I have any update. Any idea is...

For now, the annotation of the share-manager is not set, but it is feasible. The potential issue is the service interruption because the share-manager contains the nfs server.

cc @c3y1huang @innobead

I've run the performance benchmark. The performance enhancement is mainly in read operations because of the reused buffers. **_Read IOPS is increased by 5% and bandwidth is increased by 2_**%....

Yeah, read for testing. The PR in longhorn-engine is already merged.

It's a flaky issue. 5 tests and 4 passed. https://ci.longhorn.io/job/private/job/longhorn-tests-regression/1831/ ~ https://ci.longhorn.io/job/private/job/longhorn-tests-regression/1835/ In the failed test, the pod becomes `running`, but `exec_command_in_pod ` causes the error. ``` target_pod = \...

Also saw another failure in ``` command = 'ss -a -n | grep :8500 | wc -l' for i in range(RETRY_EXEC_COUNTS): > socket_cnt = exec_command_in_pod( core_api, command, mgr_name, 'longhorn-system') ```...

> It's a flaky issue. 5 tests and 4 passed. https://ci.longhorn.io/job/private/job/longhorn-tests-regression/1831/ ~ https://ci.longhorn.io/job/private/job/longhorn-tests-regression/1835/ > > In the failed test, the pod becomes `running`, but `exec_command_in_pod ` causes the error. >...

Did an investigation of the `engineimage.status.refcount`. The refcount is caculated in engineimage controller ([here](https://github.com/longhorn/longhorn-manager/blob/v1.2.4/controller/engine_image_controller.go#L314) and [here](https://github.com/longhorn/longhorn-manager/blob/v1.2.4/controller/engine_image_controller.go#L559)). It counts the volumes whose spec or status are using this engineimage. To prevent...