[ https://issues.apache.org/jira/browse/HDFS-1057?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12871497#action_12871497 ]
Hairong Kuang commented on HDFS-1057: ------------------------------------- Sam, the replica length in my comment meant the visible length so no new length field is needed. The read algorithm in trunk simplifies the read algorithm in the design doc. A block reader does read until its disk length >= the request # of bytes. To solve the last chunk CRC problem, I am thinking to store last chunk's crc in ReplicaBeingWritten after the chunk is received. A block reader returns visible length of bytes and last chunk crc from memory. This should work because # of requested bytes <= disk length <= visible length. > Concurrent readers hit ChecksumExceptions if following a writer to very end > of file > ----------------------------------------------------------------------------------- > > Key: HDFS-1057 > URL: https://issues.apache.org/jira/browse/HDFS-1057 > Project: Hadoop HDFS > Issue Type: Sub-task > Components: data-node > Affects Versions: 0.21.0, 0.22.0 > Reporter: Todd Lipcon > Assignee: sam rash > Priority: Blocker > Attachments: conurrent-reader-patch-1.txt, > conurrent-reader-patch-2.txt, conurrent-reader-patch-3.txt > > > In BlockReceiver.receivePacket, it calls replicaInfo.setBytesOnDisk before > calling flush(). Therefore, if there is a concurrent reader, it's possible to > race here - the reader will see the new length while those bytes are still in > the buffers of BlockReceiver. Thus the client will potentially see checksum > errors or EOFs. Additionally, the last checksum chunk of the file is made > accessible to readers even though it is not stable. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.