!
Best regards,
Wei
---------
Wei Tan, PhD
Research Staff Member
IBM T. J. Watson Research Center
http://researcher.ibm.com/person/us-wtan
From: Zesheng Wu
To: user@hbase.apache.org,
Date: 03/13/2014 09:33 PM
Subject:Re: Occasional GSSException that brings down region server
Thank you al
ese errors on this page seem to be for
"does not work at all" conditions... not a "fails every randomly long
amount of time"
3) we don't have this "problematic combination of components" listed...
but again - this is a work / no work dichotomy...
Thanks,
W
,
Wei
-----
Wei Tan, PhD
Research Staff Member
IBM T. J. Watson Research Center
http://researcher.ibm.com/person/us-wtan
-
Wei Tan, PhD
Research Staff Member
IBM T. J. Watson Research Center
http://researcher.ibm.com/person/us-wtan
Image :)
Best regards,
Wei
-
Wei Tan, PhD
Research Staff Member
IBM T. J. Watson Research Center
http://researcher.ibm.com/person/us-wtan
From: Vladimir Rodionov
To: "user@hbase.apache.org" ,
Date: 02/27/2014 01:22 AM
Subject:RE: I
I am thinking of storing medium sized objects (~1M) using HBase. The
advantage of using HBase rather than HBase (storing pointers) + HDFS, in
my mind, is:
data locality. When I want to run analytics, I will access these objects
using HBase scan, and HBase stores KVs in a sequential manner. If I
Thanks for the help from you all.
Best regards,
Wei
---------
Wei Tan, PhD
Research Staff Member
IBM T. J. Watson Research Center
http://researcher.ibm.com/person/us-wtan
From: Ted Yu
To: "user@hbase.apache.org" ,
Date: 01/29/2014 04:37 PM
Subject:
criteria.
Best regards,
Wei
-
Wei Tan, PhD
Research Staff Member
IBM T. J. Watson Research Center
http://researcher.ibm.com/person/us-wtan
From: lars hofhansl
To: "user@hbase.apache.org" ,
Date: 01/29/2014 03:49 PM
Subject:Re: larger H
To be more clear, each KV (cell) is a couple of KB but each row is a
couple of MB. If I need to search through row key, but always fetch rows
as a whole, shall I use a block size larger than the default 64KB?
Thanks,
Wei
-
Wei Tan, PhD
Research Staff Member
IBM
Hi, I have a HBase table where each row has ~1000k columns, ~2K each. My
table scan pattern is to use a row key filter but I need to fetch the
whole row (~1000 k) columns back.
Shall I set HFile block size to be larger than the default 64K?
Thanks,
Wei
-
Wei Tan
Roman, thanks for sharing your experience. Is your approach is somewhat
similar to Facebook's image store, Haystack?
I am very interested in knowing your use case, and what you actually mean
by class abstraction, internal write buffer, etc.
Best regards,
Wei
From: Roman Nikitchenko
To:
Hi Gary, thanks for your clarification and yes, I totally agree with your
statement.
The class is not removed but the CP is kind of removed and not active
after an un-handled exception.
I will take a look at the Jira you mentioned.
Best regards,
Wei
From: Gary Helmling
To: user@hbase.
Hi Gary, thanks!
It seems that the region observer been removed behavior, is per region and
NOT per coprocessor. So do I have to query each region to get the per
region health status? Or, is there a table level API telling me something
like, I have 10 regions and an observer has been removed in 2
Hello, I want to check the status of each coprocessor, in a given table.
Let's say I have 3 CPs and one of them is removed due to some unhandled
exception, so I want to see this status (3 deployed, 2 currently alive).
I found this from
http://blogs.apache.org/hbase/entry/coprocessor_introductio
We would like to avoid such interruption. A global hashmap storing such a
setting would be more desirable.
Thanks,
Wei
-
Wei Tan, PhD
Research Staff Member
IBM T. J. Watson Research Center
http://researcher.ibm.com/person/us-wtan
From: Vladimir Rodionov
To
RE: passing a parameter to an observer coprocessor
>>
>> Add fake cf + column to your Put operation
>>
>> Put put = new Put(row);
>>
>> put.addFamily("COMMAND".getBytes(),
"flag".getBytes(),"true".getBytes());
>>
>> Be
Hi all,
I want to add some dynamic behavior to my observer cp, say:
postPut(){
if(flag) {do function1()};
else {do function2()}
}
Is there a way to dynamically change the value of flag? One Feasible
approaches is to change a value in table descriptor, but then I need to
restart the tab
or data processing can be seen as
>>> "unnecessary overhead".
>>>
>>> Thanks
>>>
>>> yong
>>> On Mon, Feb 18, 2013 at 10:35 AM, Michael Segel
>>> wrote:
>>>> Why?
>>>>
&g
n a lot of cases, RPC may NOT be the bottle neck. Remember a "local"
put also involves RPC -- during WAL to HDFS.
Best Regards,
Wei
Wei Tan
Research Staff Member
IBM T. J. Watson Research Center
Yorktown Heights, NY 10598
w...@us.ibm.com; 914-945-4386
From: Prakash Kadel
Is your CheckAndPut involving a local or remote READ? Due to the nature of
LSM, read is much slower compared to a write...
Best Regards,
Wei
From: Prakash Kadel
To: "user@hbase.apache.org" ,
Date: 02/17/2013 07:49 PM
Subject:coprocessor enabled put very slow, help please~~~
We need to parse JSON in a coprocessor and if HBase /lib directory
contains any json processing util, we can avoid introducing additional
jars.
Thanks!
Best Regards,
Wei
Thanks. I may try this approach later.
Now I am using a remote cluster to test and I have this workaround:
mvn install and skip test
copy the cp to the remote cluster
mvn install with test
I doubt if it is a good approach but it works.
Best Regards,
Wei
From: Adrien Mogenet
To: user@
Hi,
I am writing a maven junit test for a HBase coprocessor. The problem is
that, I want to write a junit test that deploy the cp jar into a cluster,
and test its function. However, test is before install so I cannot get a
cp jar to deploy at that time.
Is this like a chicken-and-egg problem? An
Hi Ted, Thanks for your remind. I made it in two steps:
1. write a java problem to create a table using original table's
descriptor.
2. use ./bin/hbase org.apache.hadoop.hbase.mapreduce.CopyTable to copy the
data
Best Regards,
Wei
Wei Tan
Research Staff Member
IBM T. J. Watson Res
Hi I have a table in cluster 1 with quite lengthy metadata (aka,
descriptor). How can I replicate the data as well as descriptor to another
table?
CopyTable, Export/Import does not handle table descriptor.
Thanks,
Wei
Hi,
I did a quick lookup and found these two threads related to
FuzzyRowFilter status:
https://issues.apache.org/jira/browse/HBASE-7628
https://issues.apache.org/jira/browse/HBASE-6509
It seems that FuzzyRowFilter will be available on 96 and there is a
plan to port it to 94.5. So that is th
avoid RPC.
Otherwise, if you can tolerate consistency but not latency, you put RPCs
in a queue and process them in the background. By this means you can have
a "global" index with some lag.
Best Regards,
Wei
Wei Tan
Research Staff Member
IBM T. J. Watson Research Center
Yorktown
Andrew, could you explain more, why doing cross-table operation is an
anti-pattern of using CP?
Durability might be an issue, as far as I understand. Thanks,
Best Regards,
Wei
From: Andrew Purtell
To: "user@hbase.apache.org" ,
Date: 01/12/2013 09:39 PM
Subject:Re: Coprocess
Hmm... I appreciate it if anybody can point me how to measure server side
Get/Scan time... Thanks!
Best Regards,
Wei
From: Wei Tan/Watson/IBM
To: hbase-u...@hadoop.apache.org,
Date: 12/02/2012 12:21 PM
Subject:Meaure server time of Get/Scan - through RPC logging?
Hi I am
Hi I am using YCSB to measure the performance of some complex Get/Scan,
i.e., with some filters. I found that I got weird high latency from YCSB
client but both HBase server and YCSB client are NOT saturated (in terms
of network, I/O, CPU). So I want to measure the server time a single
Get/Scan
Actually coprocessor can be used to implement MR-like function, while not
using Hadoop framework.
Best Regards,
Wei
Wei Tan
Research Staff Member
IBM T. J. Watson Research Center
Yorktown Heights, NY 10598
w...@us.ibm.com; 914-784-6752
From: Dalia Sobhy
To: "user@hbase.apach
I wonder if there is any follow up on this issue, i.e., a put can return a
timestamp of the record? Thanks!
Best Regards,
Wei
From: Wei Tan/Watson/IBM
To: user@hbase.apache.org,
Date: 08/02/2012 12:37 PM
Subject:RE: Retrieve Put timestamp
+1.
So far I think timestamp is
Hi,
I am monitoring the readRequestsCount shown in the "Requests" column in
the web GUI of a server/region. I observe that, while a put correspond to
ONE write request, a get corresponds to 2 readRequestsCount. Is that true
and is there a reason for that? I got the same number in a table with
anted overhead.
-Anoop-
________
From: Wei Tan [w...@us.ibm.com]
Sent: Wednesday, October 10, 2012 2:05 AM
To: hbase-u...@hadoop.apache.org
Subject: connect to the region from coprocessor
Hi,
In my preput coprocessor I would like to get the old value of the row been
input. Now I am creatin
Hi,
In my preput coprocessor I would like to get the old value of the row been
input. Now I am creating a HTable instance and using the get interface;
function wise it works fine. Given the row is physically in the same
region as the cp, is there any lightweight approach doing that?
Thanks!
for the chained put (seems to be an index put) to be completed -- this is
causing the timeout.
Best Regards,
Wei
Wei Tan
Research Staff Member
IBM T. J. Watson Research Center
19 Skyline Dr, Hawthorne, NY 10532
w...@us.ibm.com; 914-784-6752
From: Henry JunYoung KIM
To: user
oposed by Prof. Eric Brewer...
http://en.wikipedia.org/wiki/Eric_Brewer_%28scientist%29
Best Regards,
Wei
Wei Tan
Research Staff Member
IBM T. J. Watson Research Center
19 Skyline Dr, Hawthorne, NY 10532
w...@us.ibm.com; 914-784-6752
From: Lin Ma
To: user@hbase.apache.org,
Date:
.
Thanks,
Best Regards,
Wei
Wei Tan
Research Staff Member
IBM T. J. Watson Research Center
19 Skyline Dr, Hawthorne, NY 10532
w...@us.ibm.com; 914-784-6752
From: "Amlan Roy"
To: ,
Date: 08/06/2012 09:05 AM
Subject:RE: HBase MapReduce - Using mutiple tables as so
+1.
So far I think timestamp is very useful. I would imagine if we can
configure the return, say in pre/post put, it would be even nicer.
Thanks,
Wei
Wei Tan
Research Staff Member
IBM T. J. Watson Research Center
19 Skyline Dr, Hawthorne, NY 10532
w...@us.ibm.com; 914-784-6752
From
We have a similar requirement and here is the solution in our mind:
add a coprocessor, in prePut() get the current ms and set it to put ---
the current implementation get the current ms and set it in put()
return the ms generated to prePut() to client. For now put() does not
return any value. we
40 matches
Mail list logo
