FB chat runs on it, and also some HBase installations.
--
Alexander Lorenz
http://mapredit.blogspot.com
On Jan 12, 2012, at 1:49 PM, Jonathan Disher wrote:
> Has anyone had any experience building high-performance datanodes using SSD
> for storage? Any gotchas or notes to contribute? We were
try this:
fuse_dfs#dfs://NAMENODE:PORT /mnt fuse usetrash,rw 0 0
--
Alexander Lorenz
http://mapredit.blogspot.com
On Jan 12, 2012, at 5:05 AM, Stuti Awasthi wrote:
> Hi,
> I modified the /etc/fstab to following :
> fuse_dfs#dfs://slave:54310 /mnt fuse allow_other,rw,usetrash 0 0
>
> Now I am ju
ains Hadoop*.jar, and $JAVA_HOME/jre/lib/*jar files.
> JAVA_HOME is set differently also
> What I am doing wrong. Any Idea ??
>
> Thanks
>
>
> -Original Message-
> From: alo.alt [mailto:wget.n...@googlemail.com]
> Sent: Saturday, January 07, 2012 11:20 PM
> To: hd
(Success), outsize: 40
> unique: 2, opcode: GETATTR (3), nodeid: 1, insize: 56
> Error occurred during initialization of VM
> java/lang/NoClassDefFoundError: java/lang/Object
>
> -Original Message-
> From: alo.alt [mailto:wget.n...@googlemail.com]
> Sent: Friday, Jan
es, block locations are returned to the client in the
> order of its closeness to the reader. The block is read from data nodes in
> this preference order."
>
> If two datanodes have equal closeness, I'd like to know how the NameNode
> chooses between them.
>
&g
Ben,
the scenario should not happen, if one DN has 20 clients and the other zero
(same block) the cluster (or DN) has another problem. Rack Awareness is
described here:
https://issues.apache.org/jira/secure/attachment/12345251/Rack_aware_HDFS_proposal.pdf
- Alex
--
Alexander Lorenz
http://mapr
Stuti, define in CLASSPATH="…." only the jars you really need for. An export of
all jars in a given directory is a red flag (done with *.jar).
- Alex
On Jan 6, 2012, at 7:23 AM, M. C. Srivas wrote:
>
> unique: 1, error: 0 (Success), outsize: 40
> unique: 2, opcode: GETATTR (3), nodeid: 1, i
