[vitastor-osd] Failed to connect to peer OSD

[lnsyyj]

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

[vitalif]

Is it still the case?

[vitalif]

I'll have to check it probably if so :)

[lnsyyj]

Yes, this should happen 100%.

[vitalif]

Fixed in master branch, will go into 0.6.9