Open lnsyyj opened 3 years ago
Hi @vitalif ,
For a 3-node etcd cluster, when the virtual machine is reading and writing, stop two etcd, and then start the two etcd. There is no way for a virtual machine to automatically recover IO, and it can only be restarted.
qemu log: https://github.com/lnsyyj/vitastor-log/tree/main/Failed_to_connect_to_peer_OSD
Is it still the case?
I'll have to check it probably if so :)
Yes, this should happen 100%.
Fixed in master branch, will go into 0.6.9
Hi @vitalif ,
For a 3-node etcd cluster, when the virtual machine is reading and writing, stop two etcd, and then start the two etcd. There is no way for a virtual machine to automatically recover IO, and it can only be restarted.
qemu log: https://github.com/lnsyyj/vitastor-log/tree/main/Failed_to_connect_to_peer_OSD