OK, problem solved, hopefully. I reset the read-only flag on all indices with
curl -X PUT “localhost:9200/_all/_settings?pretty” -H ‘Content-Type: application/json’ -d’
{
“index.blocks.read_only_allow_delete”: null
}
’
And that appears to have done the trick. From what I gather, ES put all indices into read-only mode when the system’s disk space fell below the low watermark.