| 摘要 |
In one embodiment, a system comprises a plurality of nodes. On each node, a storage virtualization component may receive the I/O operations generated from local applications on the node. The storage virtualization component may transmit the I/O operation, and detect any failures that occur in the I/O operations. If a failure is detected, the storage virtualization component may cooperate with other storage virtualization components on other nodes to successfully complete an I/O operation that experiences a failure. In one embodiment, if the I/O operation is successful on a remote node, the storage virtualization component may migrate the application that sourced the I/O operation to the remote node. In other embodiments, the storage virtualization component may monitor I/O operations and migrate the application after a pattern of local failures and successes on the remote node. |
