From 4736db1b1caca50a83bf7fb0d45e2e8d48d4e233 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Mon, 10 Dec 2018 13:01:00 +0800
Subject: crank hbase GROBID worker memory usage down

---
 python/kafka_grobid_hbase.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'python')

diff --git a/python/kafka_grobid_hbase.py b/python/kafka_grobid_hbase.py
index b6219eb..b52c386 100755
--- a/python/kafka_grobid_hbase.py
+++ b/python/kafka_grobid_hbase.py
@@ -146,7 +146,7 @@ class KafkaGrobidHbaseWorker:
             managed=True,
             auto_commit_enable=True,
             # needed to avoid MessageSet decode errors
-            fetch_message_max_bytes=32*1024*1024,
+            fetch_message_max_bytes=4*1024*1024,
             # LATEST because best to miss processing than waste time re-process
             auto_offset_reset=pykafka.common.OffsetType.LATEST,
             compacted_topic=True)
-- 
cgit v1.2.3