Minor correction/addition: "Stargate" is undergoing shared development in two
github trees:
http://github.com/macdiesel/stargate/tree/master
http://github.com/apurtell/stargate/tree/master
The github network view can show which branch is leading at any point in time:
http://github.com/
Is there a requirement for hadoop 0.20 for HBase 0.20?
On Wed, Jun 17, 2009 at 1:44 AM, Andrew Purtell wrote:
> Minor correction/addition: "Stargate" is undergoing shared development in
> two github trees:
> http://github.com/macdiesel/stargate/tree/master
> http://github.com/apurtell/s
If you follow the "Getting Started" link that Stack gave, in the
"Requirements" section, you can see that this version of HBase will only run
on Hadoop 0.20
J-D
On Wed, Jun 17, 2009 at 7:32 AM, jason hadoop wrote:
> Is there a requirement for hadoop 0.20 for HBase 0.20?
>
>
> On Wed, Jun 17, 200
Hi all,
I'm running HBase 0.19.3 with Hadoop 0.19.1 on a clusters of 2 machines
operating Linux Ubuntu. Files are not being stored inside the /tmp folder.
The problem that already occurred 3 times is that, suddenly, all data stored
in my table is gone after the entire cluster is restarted, either
Hi Lucas!
Just a quick thought. Do you have a lot of data in your cluster or just a
few things in there?
If you don't have that much data in HBase it might not have been flushed to
disk/HDFS yet
and therefore only sits in the internal memcache in HBase, so when your
machines are turned
of, that dat
Hi Erik,
I have only a small amount of data, something between 1500 e 3000 documents.
Is there a way to force a flush of those documents?
1500 to 3000 is the number of new documents that the application I'm
currently working on inserts everyday, so I think it would be nice to flush
them all to di
But isn't it strange that the whole table suddenly became unavailable?
Specially because it's inside HDFS.
Also, I've already created tables with very few rows, 250 for instance, that
kept available after shutting down and starting again HBase. Is it because
when HBase is properly shut data is flu
Hi Lucas!
Yeah, have a look at HBaseAdmin and you will find flush and compact. Not
sure that compact is going to make
a big difference in your case, since you only have one flush or so per day,
but might be nice for you to run it too.
Running a compaction means that all you flushed files will be re
Lucas,
Your table is "missing" because the edits in the META table aren't flush, in
0.20 we "fix" this by setting a very small maximum memcache size on both
ROOT and META tables so that the edits go to disk often. If all the nodes in
your cluster are shutdown at the same moment, another problem th
Hi Lucas!
Not sure if you have had a look at the BigTable paper, link in the beginning
of http://hadoop.apache.org/hbase/ might clear some of the confusion.
But basically what happens is to support fast writes we only write to
memory and periodically flush this data to disk, so while data is still
Helped a lot! Thanks for the replies. I'll keep coding and move to newer
versions of HBase and Hadoop as soon as they are out. I'll also have a look
at the flush operation from HBaseAdmin.
Lucas
On Wed, Jun 17, 2009 at 1:58 PM, Erik Holstad wrote:
> Hi Lucas!
> Not sure if you have had a look
See also 'tools' in the hbase shell. There is a tool to flush all in a
table or an individual region.
I also need to roll a 0.19.4 candidate. It has a few issues that have us
flushing catalog tables way more frequently that we used to.
St.Ack
On Wed, Jun 17, 2009 at 10:04 AM, Lucas Nazário dos
Sweet thanks stack. I'll be upgrading as well. My client program takes far
too long to simply open a scanner. This problem appears to have been
addressed in .20 (https://issues.apache.org/jira/browse/HBASE-1118).
In order to skip reloading the data I do the following:
1. Shutdown hadoop/hbas
In the original email:
" In particular, this alpha release is without a migration
script to bring your 0.19.x era data forward to work on hbase 0.20.0. "
There is no migration for 0.19.x datafiles yet. That is due for the RC a
few weeks away.
If you can reload your data, that'd be best for now.
Opps missed that bit. Okay, thanks Ryan.
Ryan Rawson wrote:
>
> In the original email:
>
> " In particular, this alpha release is without a migration
> script to bring your 0.19.x era data forward to work on hbase 0.20.0. "
>
> There is no migration for 0.19.x datafiles yet. That is due fo
On Wed, Jun 17, 2009 at 2:34 PM, llpind wrote:
>
> Sweet thanks stack. I'll be upgrading as well. My client program takes
> far
> too long to simply open a scanner. This problem appears to have been
> addressed in .20 (https://issues.apache.org/jira/browse/HBASE-1118).
Or was it HBASE-867?
In the NOSQL meetup slides the inserts and reads are really good, but the
test is on single column and only 16bytes,
I wonder how the numbers would be affected if the row grows to 1K bytes,
even 16Kbytes?
if the numbers are disk I/O bounded, then we almost have to multiply the
numbers by 64 or 102
Hey,
The interesting thing is due to the way things are handled internally, small
values are more challenging than large ones. The performance is not
strictly IO bound or limited, and you won't be seeing corresponding
slowdowns on larger values.
I encourage you to give download the alpha and giv
And when I say 'test suite' i really mean "performance suite" -- that's the
problem, test suites we've been running test the functionality, not the
speed in a repeatable/scientific manner.
-ryan
On Wed, Jun 17, 2009 at 5:46 PM, Ryan Rawson wrote:
> Hey,
>
> The interesting thing is due to the
Hmmm, don't we have a performance benchmark for comparing with Bigtable?
seems a while since someone updates that...
I was just hoping that someone has a rough number in mind, so that i don't
get any big surpirse when i try this out on the larger row size data.
Thanks!
On Wed, Jun 17, 2009 at 5:5
>From the talk given at hadoop summit:
Fat Table: 1000 Rows with 10 Columns,1MB values
Sequential insert – 68 seconds (68 ms/row)
Random reads – 56.92 ms/row (average)
Full scan – 35 seconds (3.53 seconds/100 rows, 35ms/row)
so for 1 MB values, we are getting a value in 56ms. Scans in 35ms/row
That's really cool~
Thanks for the info, Ryan!!!
Cheers,
Ski Gh
On Wed, Jun 17, 2009 at 5:59 PM, Ryan Rawson wrote:
> From the talk given at hadoop summit:
>
> Fat Table: 1000 Rows with 10 Columns,1MB values
> Sequential insert – 68 seconds (68 ms/row)
> Random reads – 56.92 ms/row (average)
>
22 matches
Mail list logo