hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Zhangyong (P)" <zhangyongos.zhangy...@huawei.com>
Subject synchronized or readwritelock in on datanode
Date Fri, 13 Feb 2015 01:34:11 GMT
Hi guys,

Too many concurrent read and write on our hdfs cluster, we face input and output stream are
abnormal and block lost (it is on disk but not in block report).
We find many synchronized in FsDatasetImpl on many method, is it better to use readwritelock?


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message