Shard allocated for local recovery (post api), should exists, but doesn't

We're getting the following log message for one of our indices, and are
unable to recover one of its shards:

[2013-04-08 14:05:42,470][WARN ][cluster.action.shard ] [search-00]sending failed shard
for [messages_20130405][1], node[q2ojLu4QSIabQQs2VUl-_g], [P], s[
INITIALIZING], reason [Failed to start shard, message [
IndexShardGatewayRecoveryException[[messages_20130405][1] shard allocated
for local recovery (post api), should exists, but doesn't]]]

I see plenty of data on 3 of the nodes (2 replicas) for this shard - is it
corrupted somehow? We're on 0.20.6, so I don't think this is related
to https://github.com/elasticsearch/elasticsearch/issues/2812. Any ideas
why this may have happened, or what I might do to further troubleshoot/fix
it?

Thanks in advance,

Devon

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.