Hi!
We are currently migrating from our Elasticsearch 2.4 to 5.2.2 (or even 5.3.0).
We decided to reindex the data instead of using snapshot/restore.
For reindexing I use the Reindex API which works quite well what I have seen so far.
But for one index (so far) the reindex aborts with the following error:
{
"error": {
"root_cause": [{
"type": "parsing_exception",
"reason": "[_shards] failed to parse field [failures]",
"line": 1,
"col": 347
}],
"type": "parsing_exception",
"reason": "[search_response] failed to parse field [_shards]",
"line": 1,
"col": 347,
"caused_by": {
"type": "parsing_exception",
"reason": "[_shards] failed to parse field [failures]",
"line": 1,
"col": 347,
"caused_by": {
"type": "illegal_argument_exception",
"reason": "[failure] index doesn't support values of type: VALUE_NULL"
}
}
},
"status": 400
}
Cluster log:
The process doesn't return this error immediately.
It reindexes around 40G out of 80G and then it stops with this error.
On the one hand it seems that the response from the remote server is not valid JSON.
I saw a change in the code related to that improving the error message in 5.3.0 (https://github.com/elastic/elasticsearch/pull/22536).
But I am not sure if it just the API or if there is something wrong with a document.
There was already the case that reindex aborted because a document was not valid JSON in the 2.4 cluster.
I changed the log level to debug, but I don't know if I can see at which document it is failing.
Thanks in advance for checking the issue!