[ https://issues.apache.org/jira/browse/HDFS-7686?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14317511#comment-14317511 ]
Colin Patrick McCabe commented on HDFS-7686: -------------------------------------------- bq. Unused Iterator import in the test, LoadingCache import in VolumeScanner ok bq. I like the cache since it's a nice way of preventing scanning the same blocks over and over again, but it'd be good to also use a LinkedHashMap instead of the LinkedList and also check existence in there before adding. That way we never have dupes in the suspect queue. It seems possible to have a working set bigger than the 1000 element cache size, like if an entire disk goes bad. good idea... I'll use a {{LinkedHashSet}}. > Re-add rapid rescan of possibly corrupt block feature to the block scanner > -------------------------------------------------------------------------- > > Key: HDFS-7686 > URL: https://issues.apache.org/jira/browse/HDFS-7686 > Project: Hadoop HDFS > Issue Type: Bug > Affects Versions: 3.0.0 > Reporter: Rushabh S Shah > Assignee: Colin Patrick McCabe > Priority: Blocker > Attachments: HDFS-7686.002.patch > > > When doing a transferTo (aka sendfile operation) from the DataNode to a > client, we may hit an I/O error from the disk. If we believe this is the > case, we should be able to tell the block scanner to rescan that block soon. > The feature was originally implemented in HDFS-7548 but was removed by > HDFS-7430. We should re-add it. -- This message was sent by Atlassian JIRA (v6.3.4#6332)