5.4.1 中search线程池核心线程数没有做限制的原因

 private int applyHardSizeLimit(String name, int requestedSize) {

        int availableProcessors = EsExecutors.boundedNumberOfProcessors(settings);

        if (name.equals(Names.BULK) || name.equals(Names.INDEX)) {

            // We use a hard max size for the indexing pools, because if too many threads enter Lucene's IndexWriter, it means

            // too many segments written, too frequently, too much merging, etc:



            // TODO: I would love to be loud here (throw an exception if you ask for a too-big size), but I think this is dangerous

            // because on upgrade this setting could be in cluster state and hard for the user to correct?

            return Math.min(requestedSize, availableProcessors);

        } else {

            return requestedSize;

        }

    }