HBASE-27099 The log printing fspread/fsread cost time unit should be …

…milliseconds (#4500)

Co-authored-by: xiaozhang <issac.zhang@huolala.cn>
Signed-off-by: Duo Zhang <zhangduo@apache.org>
(cherry picked from commit b498efd)

hbase-server/src/main/java/org/apache/hadoop/hbase/io/hfile/HFileBlock.java

@@ -1705,7 +1705,7 @@ protected HFileBlock readBlockDataInternal(FSDataInputStream is, long offset,
         if (!fileContext.isCompressedOrEncrypted()) {
           hFileBlock.sanityCheckUncompressed();
         }
-        LOG.trace("Read {} in {} ns", hFileBlock, duration);
+        LOG.trace("Read {} in {} ms", hFileBlock, duration);
         // Cache next block header if we read it for the next time through here.
         if (nextBlockOnDiskSize != -1) {
           cacheNextBlockHeader(offset + hFileBlock.getOnDiskSizeWithHeader(), onDiskBlock,