You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@geode.apache.org by Catherine Johnson <cj...@pivotal.io> on 2015/06/26 17:45:40 UTC
Question regarding replicated regions and scope
If I am using a replicated region with scope "distributed-ack", and I
have a client that initiates transactions (with an actual transaction
object and a start), will this have the same impact as making the
scope on the region global? Meaning, if I do a put to a replicated
region inside of a transaction, am I protecting myself from having two
values possibly pass each other? Would this have the same effect as
"global" scope?
Thanks,
Catherine
--
Catherine Johnson | GemFire Product Management | Pivotal
+1.503.704.1933 | cjohnson@pivotal.io
Re: Question regarding replicated regions and scope
Posted by Catherine Johnson <cj...@pivotal.io>.
Thanks all!
On Fri, Jun 26, 2015 at 10:29 AM, Anilkumar Gingade <ag...@pivotal.io> wrote:
> In case of global scope, every operation is performed under
> distributed-lock...
>
> In case of transaction, the d-locks are taken at the commit time, and is
> done efficiently...
>
> -Anil.
>
>
> On Fri, Jun 26, 2015 at 10:08 AM, Swapnil Bawaskar <sb...@pivotal.io>
> wrote:
>>
>> As part of commit processing, a request is sent to grab a d-lock on the
>> key involved. So, if the two transactions try to lock the same key at the
>> same time, only one will succeed (other will get a CommitConflictException
>> and will have to retry the transaction), protecting from two values passing
>> each other.
>>
>> On Fri, Jun 26, 2015 at 8:45 AM, Catherine Johnson <cj...@pivotal.io>
>> wrote:
>>>
>>> If I am using a replicated region with scope "distributed-ack", and I
>>> have a client that initiates transactions (with an actual transaction
>>> object and a start), will this have the same impact as making the
>>> scope on the region global? Meaning, if I do a put to a replicated
>>> region inside of a transaction, am I protecting myself from having two
>>> values possibly pass each other? Would this have the same effect as
>>> "global" scope?
>>>
>>> Thanks,
>>> Catherine
>>>
>>>
>>>
>>> --
>>> Catherine Johnson | GemFire Product Management | Pivotal
>>> +1.503.704.1933 | cjohnson@pivotal.io
>>
>>
>
--
Catherine Johnson | GemFire Product Management | Pivotal
+1.503.704.1933 | cjohnson@pivotal.io
Re: Question regarding replicated regions and scope
Posted by Anilkumar Gingade <ag...@pivotal.io>.
In case of global scope, every operation is performed under
distributed-lock...
In case of transaction, the d-locks are taken at the commit time, and is
done efficiently...
-Anil.
On Fri, Jun 26, 2015 at 10:08 AM, Swapnil Bawaskar <sb...@pivotal.io>
wrote:
> As part of commit processing, a request is sent to grab a d-lock on the
> key involved. So, if the two transactions try to lock the same key at the
> same time, only one will succeed (other will get a CommitConflictException
> and will have to retry the transaction), protecting from two values passing
> each other.
>
> On Fri, Jun 26, 2015 at 8:45 AM, Catherine Johnson <cj...@pivotal.io>
> wrote:
>
>> If I am using a replicated region with scope "distributed-ack", and I
>> have a client that initiates transactions (with an actual transaction
>> object and a start), will this have the same impact as making the
>> scope on the region global? Meaning, if I do a put to a replicated
>> region inside of a transaction, am I protecting myself from having two
>> values possibly pass each other? Would this have the same effect as
>> "global" scope?
>>
>> Thanks,
>> Catherine
>>
>>
>>
>> --
>> Catherine Johnson | GemFire Product Management | Pivotal
>> +1.503.704.1933 | cjohnson@pivotal.io
>>
>
>
Re: Question regarding replicated regions and scope
Posted by Swapnil Bawaskar <sb...@pivotal.io>.
As part of commit processing, a request is sent to grab a d-lock on the key
involved. So, if the two transactions try to lock the same key at the same
time, only one will succeed (other will get a CommitConflictException and
will have to retry the transaction), protecting from two values passing
each other.
On Fri, Jun 26, 2015 at 8:45 AM, Catherine Johnson <cj...@pivotal.io>
wrote:
> If I am using a replicated region with scope "distributed-ack", and I
> have a client that initiates transactions (with an actual transaction
> object and a start), will this have the same impact as making the
> scope on the region global? Meaning, if I do a put to a replicated
> region inside of a transaction, am I protecting myself from having two
> values possibly pass each other? Would this have the same effect as
> "global" scope?
>
> Thanks,
> Catherine
>
>
>
> --
> Catherine Johnson | GemFire Product Management | Pivotal
> +1.503.704.1933 | cjohnson@pivotal.io
>