hadoop-hdfs-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Justyna Kałużka <jmkalu...@gmail.com>
Subject Adding functionality based on mover
Date Sun, 12 Jun 2016 22:21:54 GMT
Hi everyone,
As my master thesis I'm developing new functionality based on Mover
function - DFS command forceReplication which would manually move blocks
between datanodes. However I'm struggling with it in case of big files -
over 7-8 MB they seem to be initially moving (I'm checking space occupied
by datanode) but then it fails. I based my solution on TestMover pointing
which blocks and nodes should be used. With smaller files it seems to work
correctly (however not always, after restarting cluster or few minutes
after putting new file).

I'm new in this mailing list, please point me out where I should describe
my functionality further - here posting code as well or maybe on JIRA?

I'd be more than grateful to hear from you, maybe I could get in touch with
other people working on Mover?


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