Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Fixed
-
None
-
None
-
Reviewed
Description
We've seen a reducer writing 200MB to HDFS with replication = 1 spending a long time in crc calculation. In particular, it was spending 5 seconds in crc calculation out of a total of 6 for the write. I suspect that it is the java-jni border that is causing us grief.
Attachments
Attachments
Issue Links
- is related to
-
HADOOP-8617 backport pure Java CRC32 calculator changes to branch-1
- Closed
-
MAPREDUCE-782 Use PureJavaCrc32 in mapreduce spills
- Closed
- relates to
-
HADOOP-5318 Poor IO Performance due to AtomicLong operations
- Resolved
-
HADOOP-7401 Unit test TestPureJavaCRC32 warmup code warms up the not-so-important loop in PureJavaCRC32.update()
- Open
-
HADOOP-6166 Improve PureJavaCrc32
- Closed
-
HADOOP-7333 Performance improvement in PureJavaCrc32
- Closed
-
HADOOP-7337 Annotate PureJavaCrc32 as a public API
- Closed
-
HDFS-496 Use PureJavaCrc32 in HDFS
- Closed