You need to open region server ports. Client directly sends put to
appropriate region server.
вторник, 17 февраля 2015 г. пользователь Vineet Mishra написал:
> -- Forwarded message --
> From: Vineet Mishra >
> Date: Tue, Feb 17, 2015 at 12:32 PM
> Subject: Hbase not taking inserts
-- Forwarded message --
From: Vineet Mishra
Date: Tue, Feb 17, 2015 at 12:32 PM
Subject: Hbase not taking inserts from Remote Machine
To: cdh-u...@cloudera.org
Hi All,
I am trying to communicate and insert some data to my
Hbase(0.98.6-cdh5.3.0) running on Hadoop 2.5 using Hbase
Thanks everyone, I was being careless
On Tue, Feb 17, 2015 at 3:22 AM, Andrew Purtell wrote:
> cated users, but those users have to ask for special
> handling by setting a flag
> . In general a
> user
> should never submit mutations
> asking for them
>
Hi, I'm closing it in servlet.destroy. I didn't see any problems here for
months. I'm using servlet lifecycle to deal with hconnection.
вторник, 17 февраля 2015 г. пользователь Liu, Ming (HPIT-GADSC) написал:
> Hi,
>
> Thank you Serega for the helpful reply and thanks Jneidi for asking this.
> I
Hi,
I have to spend a lot of time to look into the source code of HTable,
HConnectionManager.
IMHO, it seems the document on hbase website is misleading. In the hbase online
document : http://hbase.apache.org/book.html#architecture.client . It mentioned:
Hi,
Thank you Serega for the helpful reply and thanks Jneidi for asking this. I
have similar confusion.
So Serega, when does your application finally close the HConnection? Or the
connection is NEVER closed as long as your application is running? Is it OK to
NOT close the HConnection and the a
Hi,
I was on 0.98 running PerformanceEvaluation. randomRead or sequentialRead
all show good read counts on the table regions.
filterScan shows exactly the same as you had.
Looking at the code. It could be as expected. The filter has no matching
returns, so the region server is in nextInternal
> What is hbase's philosophy in this? Does it allow some degree of data
loss?
HBase doesn't "allow" data loss, in the sense that HBase never chooses on
its own to be less than fully durable. However, our client API does allow
users to submit mutations with different durability guarantees.
The def
Hi Rams,
I don't think HBase mailing list is appropriate to search for an OCR.
Please use appropriate mailing list.
~Anil
On Mon, Feb 16, 2015 at 5:27 AM, hongbin ma wrote:
> I used to came across this: https://code.google.com/p/tesseract-ocr/
> AFAIK, OCR requires training if you want to get
Using TableInputFormat directly will have better scalability than HS2.
Better still, use TableSnapshotInputFormat to work from a snapshot (since
RDDs are immutable anyway).
-n
On Monday, February 16, 2015, Marco wrote:
> Hi,
>
> I've played with the feature to expose RDD via Thrift to enable JD
Hi Alok,
"xyz$" regex is enough to get the desired result. It will perform better
than "^.*xyz$". "^.*xyz$" is doing some unnecessary work.
Thanks,
Anil Gupta
On Wed, Feb 11, 2015 at 10:44 AM, Alok Singh wrote:
> A simple String suffix check will be definitely faster, but I doubt
> you will no
balancer said "true" and it is not disabled. Thanks again for your help.
Abe
On Mon, Feb 16, 2015 at 11:23 AM, Ted Yu wrote:
> What was the output from 'balancer' command ?
>
> Was is possible that balancer was disabled ?
>
> Cheers
>
> On Mon, Feb 16, 2015 at 8:04 AM, Abe Weinograd wrote:
>
Hi,
I've played with the feature to expose RDD via Thrift to enable JDBC
access. (Spark 1.2)
val eventsView = sqlContext.createSchemaRDD(eventSchemaRdd)
eventsView.registerTempTable("Events")
HiveThriftServer2.startWithContext(sqlContext)
This works all fine.
Now, my understanding is yo
What was the output from 'balancer' command ?
Was is possible that balancer was disabled ?
Cheers
On Mon, Feb 16, 2015 at 8:04 AM, Abe Weinograd wrote:
> Ok. I forced the balancer run and am not seeing anything after a few
> minutes. Master logs isn't showing anything. Should I look at the
Ok. I forced the balancer run and am not seeing anything after a few
minutes. Master logs isn't showing anything. Should I look at the RS ones
instead?
On Mon, Feb 16, 2015 at 11:03 AM, Ted Yu wrote:
> You should see effect in the next balancer run.
>
> Cheers
>
> On Mon, Feb 16, 2015 at 7:52
You should see effect in the next balancer run.
Cheers
On Mon, Feb 16, 2015 at 7:52 AM, Abe Weinograd wrote:
> Excellent. If i trigger the balancer, should this start to help or only for
> future region creation?
>
> Thanks,
> Abe
>
> On Mon, Feb 16, 2015 at 9:35 AM, Ted Yu wrote:
>
> > Yes. T
Excellent. If i trigger the balancer, should this start to help or only for
future region creation?
Thanks,
Abe
On Mon, Feb 16, 2015 at 9:35 AM, Ted Yu wrote:
> Yes. This setting should be modified on Master.
>
> Cheers
>
> On Mon, Feb 16, 2015 at 6:27 AM, Abe Weinograd wrote:
>
> > Thanks Ted
Any one tries this recently?
hbase org.apache.hadoop.hbase.PerformanceEvaluation --nomapred --rows=100
filterScan 3
I let it run for 30 minutes and killed it. 0 requets against the table, 649
requests against the META.
Seems to not be working anymore. Did not get a chance to look at the code
yet
Yes. This setting should be modified on Master.
Cheers
On Mon, Feb 16, 2015 at 6:27 AM, Abe Weinograd wrote:
> Thanks Ted. We are putting this in the hbase-site.xml for the Master?
>
> Abe
>
> On Fri, Feb 13, 2015 at 5:03 PM, Shahab Yunus
> wrote:
>
> > Thanks, we will try that and report bac
Hongbin:
Please go through HRegion#doMiniBatchMutation().
You will see the connection of wal.append() and the WAL sync in step 7.
Cheers
On Mon, Feb 16, 2015 at 6:05 AM, Dave Latham wrote:
> Hi Hongbin,
>
> The WAL class is used internally to the region server. Typically an HBase
> write oper
Thanks Ted. We are putting this in the hbase-site.xml for the Master?
Abe
On Fri, Feb 13, 2015 at 5:03 PM, Shahab Yunus
wrote:
> Thanks, we will try that and report back.
>
> Regards,
> Shahab
>
> On Fri, Feb 13, 2015 at 4:56 PM, Ted Yu wrote:
>
> > You can make TableSkewCostFunction more pro
Hi Hongbin,
The WAL class is used internally to the region server. Typically an HBase
write operation will first call WAL.append() with the data, then later,
after releasing locks, call WAL.sync() to ensure that the data for that
write has been synced to be durable before returning to the client
I used to came across this: https://code.google.com/p/tesseract-ocr/
AFAIK, OCR requires training if you want to get a high quality recognition.
and it's not easy to have a model that suits all styles of hand writings
On Mon, Feb 16, 2015 at 7:33 PM, N. Ramasubramanian <
ramasubramanian.naraya...@
hi, all
It seems WAL.append() in hbase, the javadoc says:
* * Append a set of edits to the WAL. The WAL is not flushed/sync'd after
this transaction*
* * completes BUT on return this edit must have its region edit/sequence
id assigned*
* * else it messes up our unification of mvcc and sequenc
Hi Andrey,
thanks for your reply, I found them very inspiring.
I'm curious about the scheme you described: It looks to me you've sharded
your data into separate regions, are you fixing the number of these regions?
If yes, each of the growing region will spawn lots of compactions while
writing da
In your scene, per servlet used a HConnecton object, so the response for 5
rpm is so easy.
If these servlets should share the only one HConnecton object, whether or not
the response latency for 5 rpm would increase?
Newrelic shows 50K RPM
each request to servlet == 1-3 put/get to HBase. I have mixed workload.
Is it strange :) ?
2015-02-16 10:37 GMT+03:00 David chen :
> 5 rpm? I am curious how the result is concluded?
27 matches
Mail list logo
