You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hbase.apache.org by Federico Gaule <fg...@despegar.com> on 2013/08/21 15:00:18 UTC
Coprocessors - failure
Hi everyone,
Let's say, I have a table(A) where every time I write a coprocessor,
hooked to postPut, writes in other table (B). I can't find anywhere what
happen if the coprocessor fails or can't connect to B
Does the client get aware of that? In case it doesn't, how can i get
noticed about the failure?
Does the Put get rolled back (I think coprocessors works outside row
transaction) ?
Thanks!
Federico
Re: Coprocessors - failure
Posted by Jung-Yup Lee <lj...@gmail.com>.
X.
ㅌ..,,(?:..;?.
2013년 8월 21일 수요일에 Ted Yu님이 작성:
> See http://search-hadoop.com/m/XtAi5Fogw32
>
>
> On Wed, Aug 21, 2013 at 6:56 AM, Federico Gaule <fgaule@despegar.com<javascript:;>>
> wrote:
>
> >
> > Table A and B are within the same HBase cluster. Can't guarantee they are
> > in the same physical machine.
> >
> > Thanks
> >
> > On 08/21/2013 10:54 AM, Ted Yu wrote:
> >
> >> Are tables A and B colocated ?
> >> Meaning, is your coprocessor making an RPC call to another server ?
> >>
> >> Cheers
> >>
> >>
> >> On Wed, Aug 21, 2013 at 6:00 AM, Federico Gaule <fgaule@despegar.com<javascript:;>
> >
> >> wrote:
> >>
> >> Hi everyone,
> >>> Let's say, I have a table(A) where every time I write a coprocessor,
> >>> hooked to postPut, writes in other table (B). I can't find anywhere
> what
> >>> happen if the coprocessor fails or can't connect to B
> >>> Does the client get aware of that? In case it doesn't, how can i get
> >>> noticed about the failure?
> >>> Does the Put get rolled back (I think coprocessors works outside row
> >>> transaction) ?
> >>>
> >>> Thanks!
> >>> Federico
> >>>
> >>>
> >>>
> >>>
> >
>
Re: Coprocessors - failure
Posted by fg...@despegar.com.
Yes, schemas are similar, both are flat-wide, 'A' has half 'B' row key, and the other half as column (hope to be clear)
I'll take a look to that class.
Cheers
Enviado desde mi BlackBerry de Personal (http://www.personal.com.ar/)
-----Original Message-----
From: Ted Yu <yu...@gmail.com>
Date: Wed, 21 Aug 2013 16:11:01
To: Federico Gaule<fg...@despegar.com>
Reply-To: user@hbase.apache.org
Cc: user@hbase.apache.org<us...@hbase.apache.org>
Subject: Re: Coprocessors - failure
Certain work is needed on the load balancer side to colocate regions of
table A with those of table B.
Do they have similar schema ?
For reference, take a look
at secondaryindex/src/main/java/org/apache/hadoop/hbase/index/SecIndexLoadBalancer.java
from HBASE-9203
Cheers
On Wed, Aug 21, 2013 at 7:09 AM, Federico Gaule <fg...@despegar.com> wrote:
> Thanks Ted.
> Basically it's recommended to avoid using RPC calls when writing from
> table A to B, having the region in the same RegionServer. There i will
> exclude network failures. Am i right?
>
>
>
> On 08/21/2013 11:02 AM, Ted Yu wrote:
>
>> See http://search-hadoop.com/m/**XtAi5Fogw32<http://search-hadoop.com/m/XtAi5Fogw32>
>>
>>
>> On Wed, Aug 21, 2013 at 6:56 AM, Federico Gaule <fg...@despegar.com>
>> wrote:
>>
>> Table A and B are within the same HBase cluster. Can't guarantee they are
>>> in the same physical machine.
>>>
>>> Thanks
>>>
>>> On 08/21/2013 10:54 AM, Ted Yu wrote:
>>>
>>> Are tables A and B colocated ?
>>>> Meaning, is your coprocessor making an RPC call to another server ?
>>>>
>>>> Cheers
>>>>
>>>>
>>>> On Wed, Aug 21, 2013 at 6:00 AM, Federico Gaule <fg...@despegar.com>
>>>> wrote:
>>>>
>>>> Hi everyone,
>>>>
>>>>> Let's say, I have a table(A) where every time I write a coprocessor,
>>>>> hooked to postPut, writes in other table (B). I can't find anywhere
>>>>> what
>>>>> happen if the coprocessor fails or can't connect to B
>>>>> Does the client get aware of that? In case it doesn't, how can i get
>>>>> noticed about the failure?
>>>>> Does the Put get rolled back (I think coprocessors works outside row
>>>>> transaction) ?
>>>>>
>>>>> Thanks!
>>>>> Federico
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>
Re: Coprocessors - failure
Posted by Ted Yu <yu...@gmail.com>.
Certain work is needed on the load balancer side to colocate regions of
table A with those of table B.
Do they have similar schema ?
For reference, take a look
at secondaryindex/src/main/java/org/apache/hadoop/hbase/index/SecIndexLoadBalancer.java
from HBASE-9203
Cheers
On Wed, Aug 21, 2013 at 7:09 AM, Federico Gaule <fg...@despegar.com> wrote:
> Thanks Ted.
> Basically it's recommended to avoid using RPC calls when writing from
> table A to B, having the region in the same RegionServer. There i will
> exclude network failures. Am i right?
>
>
>
> On 08/21/2013 11:02 AM, Ted Yu wrote:
>
>> See http://search-hadoop.com/m/**XtAi5Fogw32<http://search-hadoop.com/m/XtAi5Fogw32>
>>
>>
>> On Wed, Aug 21, 2013 at 6:56 AM, Federico Gaule <fg...@despegar.com>
>> wrote:
>>
>> Table A and B are within the same HBase cluster. Can't guarantee they are
>>> in the same physical machine.
>>>
>>> Thanks
>>>
>>> On 08/21/2013 10:54 AM, Ted Yu wrote:
>>>
>>> Are tables A and B colocated ?
>>>> Meaning, is your coprocessor making an RPC call to another server ?
>>>>
>>>> Cheers
>>>>
>>>>
>>>> On Wed, Aug 21, 2013 at 6:00 AM, Federico Gaule <fg...@despegar.com>
>>>> wrote:
>>>>
>>>> Hi everyone,
>>>>
>>>>> Let's say, I have a table(A) where every time I write a coprocessor,
>>>>> hooked to postPut, writes in other table (B). I can't find anywhere
>>>>> what
>>>>> happen if the coprocessor fails or can't connect to B
>>>>> Does the client get aware of that? In case it doesn't, how can i get
>>>>> noticed about the failure?
>>>>> Does the Put get rolled back (I think coprocessors works outside row
>>>>> transaction) ?
>>>>>
>>>>> Thanks!
>>>>> Federico
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>
Re: Coprocessors - failure
Posted by Federico Gaule <fg...@despegar.com>.
Thanks Ted.
Basically it's recommended to avoid using RPC calls when writing from
table A to B, having the region in the same RegionServer. There i will
exclude network failures. Am i right?
On 08/21/2013 11:02 AM, Ted Yu wrote:
> See http://search-hadoop.com/m/XtAi5Fogw32
>
>
> On Wed, Aug 21, 2013 at 6:56 AM, Federico Gaule <fg...@despegar.com> wrote:
>
>> Table A and B are within the same HBase cluster. Can't guarantee they are
>> in the same physical machine.
>>
>> Thanks
>>
>> On 08/21/2013 10:54 AM, Ted Yu wrote:
>>
>>> Are tables A and B colocated ?
>>> Meaning, is your coprocessor making an RPC call to another server ?
>>>
>>> Cheers
>>>
>>>
>>> On Wed, Aug 21, 2013 at 6:00 AM, Federico Gaule <fg...@despegar.com>
>>> wrote:
>>>
>>> Hi everyone,
>>>> Let's say, I have a table(A) where every time I write a coprocessor,
>>>> hooked to postPut, writes in other table (B). I can't find anywhere what
>>>> happen if the coprocessor fails or can't connect to B
>>>> Does the client get aware of that? In case it doesn't, how can i get
>>>> noticed about the failure?
>>>> Does the Put get rolled back (I think coprocessors works outside row
>>>> transaction) ?
>>>>
>>>> Thanks!
>>>> Federico
>>>>
>>>>
>>>>
>>>>
Re: Coprocessors - failure
Posted by Ted Yu <yu...@gmail.com>.
See http://search-hadoop.com/m/XtAi5Fogw32
On Wed, Aug 21, 2013 at 6:56 AM, Federico Gaule <fg...@despegar.com> wrote:
>
> Table A and B are within the same HBase cluster. Can't guarantee they are
> in the same physical machine.
>
> Thanks
>
> On 08/21/2013 10:54 AM, Ted Yu wrote:
>
>> Are tables A and B colocated ?
>> Meaning, is your coprocessor making an RPC call to another server ?
>>
>> Cheers
>>
>>
>> On Wed, Aug 21, 2013 at 6:00 AM, Federico Gaule <fg...@despegar.com>
>> wrote:
>>
>> Hi everyone,
>>> Let's say, I have a table(A) where every time I write a coprocessor,
>>> hooked to postPut, writes in other table (B). I can't find anywhere what
>>> happen if the coprocessor fails or can't connect to B
>>> Does the client get aware of that? In case it doesn't, how can i get
>>> noticed about the failure?
>>> Does the Put get rolled back (I think coprocessors works outside row
>>> transaction) ?
>>>
>>> Thanks!
>>> Federico
>>>
>>>
>>>
>>>
>
Re: Coprocessors - failure
Posted by Federico Gaule <fg...@despegar.com>.
Table A and B are within the same HBase cluster. Can't guarantee they
are in the same physical machine.
Thanks
On 08/21/2013 10:54 AM, Ted Yu wrote:
> Are tables A and B colocated ?
> Meaning, is your coprocessor making an RPC call to another server ?
>
> Cheers
>
>
> On Wed, Aug 21, 2013 at 6:00 AM, Federico Gaule <fg...@despegar.com> wrote:
>
>> Hi everyone,
>> Let's say, I have a table(A) where every time I write a coprocessor,
>> hooked to postPut, writes in other table (B). I can't find anywhere what
>> happen if the coprocessor fails or can't connect to B
>> Does the client get aware of that? In case it doesn't, how can i get
>> noticed about the failure?
>> Does the Put get rolled back (I think coprocessors works outside row
>> transaction) ?
>>
>> Thanks!
>> Federico
>>
>>
>>
Re: Coprocessors - failure
Posted by Ted Yu <yu...@gmail.com>.
Are tables A and B colocated ?
Meaning, is your coprocessor making an RPC call to another server ?
Cheers
On Wed, Aug 21, 2013 at 6:00 AM, Federico Gaule <fg...@despegar.com> wrote:
> Hi everyone,
> Let's say, I have a table(A) where every time I write a coprocessor,
> hooked to postPut, writes in other table (B). I can't find anywhere what
> happen if the coprocessor fails or can't connect to B
> Does the client get aware of that? In case it doesn't, how can i get
> noticed about the failure?
> Does the Put get rolled back (I think coprocessors works outside row
> transaction) ?
>
> Thanks!
> Federico
>
>
>