Hi Dbruba,
Thanx for your reply.
On the first part (NameNode HA and failover), our experience with NFS has
not been very good.
Is having a Db as a backing store for NameNode an option (I understand that
this may not be part of the current release 0.15.0 and would be a new
feature)?
-Taj
Hello there, we would like to make some tests with hadoop.
For the tests we would like to have a hadoop filesystem up and configured, so
using stubs and some mocks of core interfaces we can test the overall storage
functionality we're about to develop (which would be a part of map/reduce jobs
Hi,
I am using hadoop-0.13.1 version. I am getting the following error
when the system running hadoop reboots.. It is not starting the data node.
Usually I face this problem when there is hadoop version mismatch or
during hadoop version upgrade.
The datanode log has this exception.
ERROR
Text objects typically contain more bytes than are actually in use.
If you were to use the alternate constructor for ImmutableBytesWritable:
new ImmutableBytesWritable(input.getBytes(), 0, input.getLength());
the test will pass.
One more note: Relying on the default encoding being the same for
What Jim just said, but it looks to me like Text is doing the wrong
thing. When you ask it its length, it returns the byte buffer capacity
rather than how many bytes are in use. It says length is 16 but there
are only 15 characters in your test string, UTF-8'd or not.
St.Ack
Jason Grey
Did anyone try DRBD (http://www.drbd.org/) for mirroring the fsimage
and editlogs to another machine?
Another idea which would involve code changes is to go to something
like Terracotta (http://www.terracottatech.com/) essentially allowing
multiple machines simultaneously to play the role
Well, it depends on how the Text object is initialized. If it is initialized
with a String, it sets its internal length to the length of the string.
---
Jim Kellerman, Senior Engineer; Powerset
-Original Message-
From: stack [mailto:[EMAIL PROTECTED]
Sent: Wednesday, November 21,
On review, after a kick from JK, I'm mistaken in the below. I wasn't
using Text.getLength, I was doing Text.getBytes().length. Please ignore
below.
St.Ack
stack wrote:
What Jim just said, but it looks to me like Text is doing the wrong
thing. When you ask it its length, it returns the byte
We're in for the tab (limit of $200).
On 11/20/07 10:42 PM, Erich Nachbar [EMAIL PROTECTED] wrote:
Guys,
It has been almost 2 months and I'd like to propose another Bay Area
Get Together.
I thought we could try and hit Gordon Biersch in Palo Alto again
around 5pm next Fri. (Nov. 30th).
On Nov 21, 2007, at 7:21 AM, Eugeny N Dzhurinsky wrote:
For the tests we would like to have a hadoop filesystem up and
configured, so
using stubs and some mocks of core interfaces we can test the
overall storage
functionality we're about to develop (which would be a part of map/
reduce
Is there a roadmap in place to make the Namenode highly available (ignoring
scalability)? I'm curious as to the priority of high availability for the
Yahoo folks in particular.
On 11/21/07, Erich Nachbar [EMAIL PROTECTED] wrote:
Did anyone try DRBD (http://www.drbd.org/) for mirroring the
Are people interested in coming by Facebook for a quick beer beforehand? We
have a happy hour on Fridays at 5 pm right down the road from Gordon
Biersch. We can show off a few of our Hadoop applications to those
interested and then head out for
On 11/21/07, Ted Dunning [EMAIL PROTECTED] wrote:
Sure. I uploaded all my logs to my website:
http://cs.duke.edu/~kcd/hadoop-logs/
My cluster consists of hadoop07-hadoop12. hadoop07 acts as the dfs
master, mapred master, and hbase master. hadoop08-12 are dfs slaves,
mapred task servers, and hbase regionservers.
I also uploaded my
in this case the map data is large enough that in-memory merges proably had no
effect (but thanks for pointing that out). (the map.out files were about
256-512MB in size - block compressed sequencefiles).
if we could initiate the on-disk merges - that would be awesome.
i am curious whether
I think that would be really great!
If you want we could also move it over to Facebook if you have the
space (hopefully I'm not jumping the gun here!)
This way we don't have to move people and there is no confusion where
to be at what time.
I can offer to bring a keg for refreshments, so we
Reduce Jobs must wait for all maps to be done before doing any work. Why are
they started before the maps are done?
example of problem
If I am running a job and its taking up all the reduce task for all nodes
and I launch a second job and see the job priority higher then the current
running
It is most common to have fewer reduce jobs than map jobs.
Also, the reason that reduce jobs start before the map jobs complete is to
avoid idling resources when possible.
This is especially important where reduce can be done in successive passes.
Counting jobs benefit from this ENORMously.
now i can't solve this problem
--
View this message in context:
http://www.nabble.com/Generator%3A-0-records-selected-for--fetching%2C-exiting-...-tf4848287.html#a13889043
Sent from the Hadoop Users mailing list archive at Nabble.com.
18 matches
Mail list logo
