hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ted Yu <yuzhih...@gmail.com>
Subject Re: Couple notes and questions on RegionObserver's
Date Fri, 26 Jul 2013 14:33:56 GMT
There was some effort in adding support for region server-based endpoint.
Here is the JIRA:

https://issues.apache.org/jira/browse/HBASE-5982

Cheers

On Wed, Jul 24, 2013 at 10:34 PM, lars hofhansl <larsh@apache.org> wrote:

> It's because a region is the independent unit in HBase. A region is
> mobile. If you pick something larger than a region you are tying state
> together that would complicate region mobility.
>
> There are probably other ways to design this, but from this angle it makes
> sense.
> If Andy Purtell is around he probably can elaborate further.
>
> -- Lars
>
> ________________________________
> From: Vladimir Rodionov <vrodionov@carrieriq.com>
> To: "dev@hbase.apache.org" <dev@hbase.apache.org>; lars hofhansl <
> larsh@apache.org>
> Sent: Wednesday, July 24, 2013 10:19 PM
> Subject: RE: Couple notes and questions on RegionObserver's
>
>
> Lars, I know that coprocessors are per region, I just do not understand
> why:
>
> see all my 5 questions in my previous e-mails.
>
> Best regards,
> Vladimir Rodionov
> Principal Platform Engineer
> Carrier IQ, www.carrieriq.com
> e-mail: vrodionov@carrieriq.com
>
> ________________________________________
> From: lars hofhansl [larsh@apache.org]
> Sent: Wednesday, July 24, 2013 8:59 PM
> To: dev@hbase.apache.org
> Subject: Re: Couple notes and questions on RegionObserver's
>
> Yes, coprocessors are per region and instantiated when the region is
> instantiated, which is what happens at enable.
> That's by design. How else would you design it? How would you allow for
> reloading updated coprocessor classes?
>
> -- Lars
>
> ________________________________
> From: Vladimir Rodionov <vrodionov@carrieriq.com>
> To: "dev@hbase.apache.org" <dev@hbase.apache.org>
> Sent: Wednesday, July 24, 2013 8:43 PM
> Subject: RE: Couple notes and questions on RegionObserver's
>
>
> This becomes more interesting ...
>
> On region.disable (Table.disable) HBase close and stops coprocessor
> On region.enable (Table.enable) HBase CREATES new instance and then call
> its start()
>
> ... thus effectively dumping all coprocessor's context and state. So, what
> happens when
> someone want sto update CF in hbase shell:
>
> disable 'Table'
>
> alter .....
>
> enable 'Table'
>
> +> Welcome new coprocessor instance.
>
>
> Best regards,
> Vladimir Rodionov
> Principal Platform Engineer
> Carrier IQ, www.carrieriq.com
> e-mail: vrodionov@carrieriq.com
>
> ________________________________________
> From: Vladimir Rodionov
> Sent: Wednesday, July 24, 2013 8:31 PM
> To: dev@hbase.apache.org
> Subject: Couple notes and questions on RegionObserver's
>
> On design:
>
> Why is the instance per region if we specify region in every call to
> RegionObserver anyway( through RegionCoprocessorEnvironment)? The
> coprocessor instance is attached to every table region and
> receives the very same region as a parameter of every call except
> start/stop?
>
> I want one coprocessor instance per RegionServer->Table or I want one
> Instance per RegionServer. Its not feasible currently?
>
> On implementation:
> 1.  stop() is always called after postClose - this one I tested
> 2. start() is always called before preOpen() ?
>
> OK. this is annoying. The contract of start/stop becomes unclear. I though
> that we start should be called on coprocessor load and
> stop - on RS shutdown.
>
> 3. abortRequested parameter is always false. Yes, its for testing but how
> can I distinguish disabling table and RS shutdown?
>
> 4. How can I intercept (or get notified on) RS shutdown request from
> inside a coprocessor?
>
> and the last one:
>
> 5. I have one table and one region - hence one coprocessor instance. When
> I disable table I observe the following sequence of calls:
>
> preClose
> preClose
> preClose
> postClose
> stop
>
> preClose is called three times in row. Is it feature or a bug?
>
> This was tested with MiniHBaseCluster. Hbase 0.94.6.1
>
> Best regards,
> Vladimir Rodionov
> Principal Platform Engineer
> Carrier IQ, www.carrieriq.com
> e-mail: vrodionov@carrieriq.com
>
> Confidentiality Notice:  The information contained in this message,
> including any attachments hereto, may be confidential and is intended to be
> read only by the individual or entity to whom this message is addressed. If
> the reader of this message is not the intended recipient or an agent or
> designee of the intended recipient, please note that any review, use,
> disclosure or distribution of this message or its attachments, in any form,
> is strictly prohibited.  If you have received this message in error, please
> immediately notify the sender and/or Notifications@carrieriq.com and
> delete or destroy any copy of this message and its attachments.
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message