ElasticSearch UNASSIGNED索引在没有数据丢失的情况下修复

Question

无论出于何种原因，一堆指数变为UNASSIGNED。我正在寻找一种方法将它们分配给群集节点，而不会丢失任何数据。

我尝试使用以下API调用，但遗憾的是导致数据丢失（由于allow_primary）：

curl -XPOST 'localhost:9200/_cluster/reroute?pretty' -d '{
        "commands" : [ {
                "allocate" : {
                    "index" : "index-name",
                    "shard" : "0",
                    "allow_primary" : true,
                    "node" : "node-name"
                }
            }
        ]
    }'

我还在elasticsearch.log中继续获得以下条目：

[2015-03-16 11:51:12,181][DEBUG][action.search.type       ] [cluster node] All shards failed for phase: [query_fetch]
[2015-03-16 11:51:12,450][DEBUG][action.search.type       ] [cluster node] All shards failed for phase: [query_fetch]
[2015-03-16 11:51:19,349][DEBUG][action.bulk              ] [cluster node] observer: timeout notification from cluster service. timeout setting [1m], time since start [1m]
[2015-03-16 11:51:20,057][DEBUG][action.bulk              ] [cluster node] observer: timeout notification from cluster service. timeout setting [1m], time since start [1m]

任何帮助都将不胜感激。