Hi,
I do concurrent bulk requests (inserting) in 20 threads. When I look at threads all of them waiting in:
Can I somehow speedup this process?
How did you get this report? (just curious).
That does not look like the name of bulk threads used in Elasticsearch.
1 Like
I named my threads which are responsible for bulk requests.
But, all of them are in WAITING state in
org.elasticsearch.common,util.concurrent.BaseFuture$Sync.get
Of course, finally the requests are done without any exception, but to compare with an older version (1.7.1), the bulk process takes much more time in the newest version.
I run this tasks in 20 threads, I know that queue size for the bulk request is 200.
Do you know how to speed up this process (I do all steps from https://www.elastic.co/guide/en/elasticsearch/reference/6.2/tune-for-indexing-speed.html)?
I run my tasks in my own thread pull. I run VisualVM and I look on threads, why this indexing process takes too much time. I see that all of my bulk threads are in WAITING state, looks like problem with asynchronus tasks.
So this is client side then. These threads are waiting on a reply from the server. You need to see why the server is taking so long to reply, maybe you're overwhelming it.
Ho, this is not from elasticsearch so?
Check the thread_pool to see what thread are doing (/_cat/thread_pool?v&human)
If I am right, for indexing 1 thread can take care of 1 primary shard.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.