2013/8/20 Liu Yuan <namei.unix at gmail.com>: > We actually purge data of backend store, but not the cache. I am not sure if > we should purge the cache too. This depends on what we have done. If the node simply went down with a running guest a) I *do not* start the guest on any other host; insert back the node; run the guest again on the same node. In such case I need the cache to be there and to be correctly flushed. b) The case of the previous mail: vdi has been removed while node was down. Cache and data should be purged. Is sheepdog able to "understand" case 'b'? |