You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ignite.apache.org by Denis Magda <dm...@apache.org> on 2016/12/20 20:05:04 UTC
Re: Hello
Alexander,
CC-ing the dev list to keep everyone in track.
I hope the case you’re talking about can be already implemented by the usage of RendezvousAffinityFunction.setAffinityBackupFilter(…) method [1]. This method gives you a way to control a locality of your primary and backup nodes. Depending on the filter implementation specific nodes might end up in one physical location while the rest in the others.
[1] https://ignite.apache.org/releases/1.7.0/javadoc/org/apache/ignite/cache/affinity/rendezvous/RendezvousAffinityFunction.html <https://ignite.apache.org/releases/1.7.0/javadoc/org/apache/ignite/cache/affinity/rendezvous/RendezvousAffinityFunction.html>
—
Denis
> On Dec 20, 2016, at 7:02 AM, Александр Меньшиков <sh...@gmail.com> wrote:
>
> I think in near time i will take some "newbie" tickets, but for future want to ask about ring topology in Ignite. In case where we have big physically stretched claster ring can be not optimal. For example: Node-1 in 1-physical location, Node-2 in 2-physical location, Node-3 in 1-physical location, Node-4 in 2-physical location and ect. So if one physical location to fall out of the cluster, we get more breaks in ring than the theoretical minimum. I want to add an additional order in the process of constructing a ring so that ensure physically close nodes were close in topology.
>
> 2016-11-24 0:01 GMT+03:00 Denis Magda <dmagda@apache.org <ma...@apache.org>>:
> Here are some ticket you may want to start with
> https://ignite.apache.org/community/contribute.html#pick-ticket <https://ignite.apache.org/community/contribute.html#pick-ticket>
>
> —
> Denis
>
>> On Nov 23, 2016, at 5:33 AM, Dmitriy Setrakyan <dsetrakyan@apache.org <ma...@apache.org>> wrote:
>>
>> Hi Alex,
>>
>> Welcome to the Ignite community!
>>
>> Once you are familiar with the contribution process, please see if there is
>> a specific Jira ticket you would like work on. We will help you get started.
>>
>> D.
>>
>> On Wed, Nov 23, 2016 at 2:24 AM, Vladislav Pyatkov <vpyatkov@gridgain.com <ma...@gridgain.com>>
>> wrote:
>>
>>> Please read https://ignite.apache.org/community/contribute.html#contribute <https://ignite.apache.org/community/contribute.html#contribute>
>>>
>>> Looking forward for your contributions!
>>>
>>> On Wed, Nov 23, 2016 at 1:15 PM, Александр Меньшиков <sharplermc@gmail.com <ma...@gmail.com>
>>>>
>>> wrote:
>>>
>>>> I want to contribute.
>>>>
>>>> 2016-11-23 13:04 GMT+03:00 Vladislav Pyatkov <vpyatkov@gridgain.com <ma...@gridgain.com>>:
>>>>
>>>>> Hi Alex,
>>>>>
>>>>> This is Apache Ignite development community.
>>>>> Are you want to contribute to Ignite or you are have any suggestion?
>>>>>
>>>>> On Wed, Nov 23, 2016 at 12:41 PM, Александр Меньшиков <
>>>>> sharplermc@gmail.com <ma...@gmail.com>>
>>>>> wrote:
>>>>>
>>>>>> Hello
>>>>>>
>>>>>
>>>>
>>>
>
>
Re: Hello
Posted by Denis Magda <dm...@apache.org>.
Alexander,
Thanks for the details. Now I see what you’re talking about. Agree, it makes sense to support this. Could you initiate a dedicated discussion on the dev list proposing this to the community and sharing design/implementations details if any? This discussion named “Hello” is not the place for this :)
—
Denis
> On Dec 21, 2016, at 1:42 AM, Александр Меньшиков <sh...@gmail.com> wrote:
>
> Denis,
>
> Thanks for your reply.
>
> My question is not about data locality. As far as I know nodes are connected in a ring. For example if i have 6 nodes, with names A, B, C, D, E, and F they can connect in ring any possible way: A-B-C-D-E-F-A, or A-F-B-E-C-D-A, etc. And if some node falls out of topology neighboring nodes must reconnect. If nodes A,B and C located in the same physical location, and D, E and F in another, and in some time one physical location is not available in another, we can get different number of reconnections. Best case scenario if we have ring like A-B-CxD-E-FxA ('x' mean disconnect) -- then we get only one reconnect (C reconnect to A or F reconnect to D -- depending on what part of the cluster we leave alive). But now possible that case AxFxBxExCxDxA -- then we get a lot of reconnections (A to B, B to C, C to A -- in general n/2 reconnections, where n -- number of nodes). And i think to add something to ensure that we always have good sorting of nodes connections (A-B-C-...-Z-A).
>
> 2016-12-20 23:05 GMT+03:00 Denis Magda <dmagda@apache.org <ma...@apache.org>>:
> Alexander,
>
> CC-ing the dev list to keep everyone in track.
>
> I hope the case you’re talking about can be already implemented by the usage of RendezvousAffinityFunction.setAffinityBackupFilter(…) method [1]. This method gives you a way to control a locality of your primary and backup nodes. Depending on the filter implementation specific nodes might end up in one physical location while the rest in the others.
>
> [1] https://ignite.apache.org/releases/1.7.0/javadoc/org/apache/ignite/cache/affinity/rendezvous/RendezvousAffinityFunction.html <https://ignite.apache.org/releases/1.7.0/javadoc/org/apache/ignite/cache/affinity/rendezvous/RendezvousAffinityFunction.html>
>
> —
> Denis
>
>> On Dec 20, 2016, at 7:02 AM, Александр Меньшиков <sharplermc@gmail.com <ma...@gmail.com>> wrote:
>>
>> I think in near time i will take some "newbie" tickets, but for future want to ask about ring topology in Ignite. In case where we have big physically stretched claster ring can be not optimal. For example: Node-1 in 1-physical location, Node-2 in 2-physical location, Node-3 in 1-physical location, Node-4 in 2-physical location and ect. So if one physical location to fall out of the cluster, we get more breaks in ring than the theoretical minimum. I want to add an additional order in the process of constructing a ring so that ensure physically close nodes were close in topology.
>>
>> 2016-11-24 0:01 GMT+03:00 Denis Magda <dmagda@apache.org <ma...@apache.org>>:
>> Here are some ticket you may want to start with
>> https://ignite.apache.org/community/contribute.html#pick-ticket <https://ignite.apache.org/community/contribute.html#pick-ticket>
>>
>> —
>> Denis
>>
>>> On Nov 23, 2016, at 5:33 AM, Dmitriy Setrakyan <dsetrakyan@apache.org <ma...@apache.org>> wrote:
>>>
>>> Hi Alex,
>>>
>>> Welcome to the Ignite community!
>>>
>>> Once you are familiar with the contribution process, please see if there is
>>> a specific Jira ticket you would like work on. We will help you get started.
>>>
>>> D.
>>>
>>> On Wed, Nov 23, 2016 at 2:24 AM, Vladislav Pyatkov <vpyatkov@gridgain.com <ma...@gridgain.com>>
>>> wrote:
>>>
>>>> Please read https://ignite.apache.org/community/contribute.html#contribute <https://ignite.apache.org/community/contribute.html#contribute>
>>>>
>>>> Looking forward for your contributions!
>>>>
>>>> On Wed, Nov 23, 2016 at 1:15 PM, Александр Меньшиков <sharplermc@gmail.com <ma...@gmail.com>
>>>>>
>>>> wrote:
>>>>
>>>>> I want to contribute.
>>>>>
>>>>> 2016-11-23 13:04 GMT+03:00 Vladislav Pyatkov <vpyatkov@gridgain.com <ma...@gridgain.com>>:
>>>>>
>>>>>> Hi Alex,
>>>>>>
>>>>>> This is Apache Ignite development community.
>>>>>> Are you want to contribute to Ignite or you are have any suggestion?
>>>>>>
>>>>>> On Wed, Nov 23, 2016 at 12:41 PM, Александр Меньшиков <
>>>>>> sharplermc@gmail.com <ma...@gmail.com>>
>>>>>> wrote:
>>>>>>
>>>>>>> Hello
>>>>>>>
>>>>>>
>>>>>
>>>>
>>
>>
>
>
Re: Hello
Posted by Александр Меньшиков <sh...@gmail.com>.
Denis,
Thanks for your reply.
My question is not about data locality. As far as I know nodes are
connected in a ring. For example if i have 6 nodes, with names A, B, C, D,
E, and F they can connect in ring any possible way: A-B-C-D-E-F-A, or
A-F-B-E-C-D-A, etc. And if some node falls out of topology neighboring
nodes must reconnect. If nodes A,B and C located in the same physical
location, and D, E and F in another, and in some time one physical location
is not available in another, we can get different number of reconnections.
Best case scenario if we have ring like A-B-CxD-E-FxA ('x' mean disconnect)
-- then we get only one reconnect (C reconnect to A or F reconnect to D --
depending on what part of the cluster we leave alive). But now possible
that case AxFxBxExCxDxA -- then we get a lot of reconnections (A to B, B to
C, C to A -- in general n/2 reconnections, where n -- number of nodes). And
i think to add something to ensure that we always have good sorting of
nodes connections (A-B-C-...-Z-A).
2016-12-20 23:05 GMT+03:00 Denis Magda <dm...@apache.org>:
> Alexander,
>
> CC-ing the dev list to keep everyone in track.
>
> I hope the case you’re talking about can be already implemented by the
> usage of RendezvousAffinityFunction.setAffinityBackupFilter(…) method
> [1]. This method gives you a way to control a locality of your primary and
> backup nodes. Depending on the filter implementation specific nodes might
> end up in one physical location while the rest in the others.
>
> [1] https://ignite.apache.org/releases/1.7.0/javadoc/org/
> apache/ignite/cache/affinity/rendezvous/RendezvousAffinityFunction.html
>
> —
> Denis
>
> On Dec 20, 2016, at 7:02 AM, Александр Меньшиков <sh...@gmail.com>
> wrote:
>
> I think in near time i will take some "newbie" tickets, but for future
> want to ask about ring topology in Ignite. In case where we have big
> physically stretched claster ring can be not optimal. For example: Node-1
> in 1-physical location, Node-2 in 2-physical location, Node-3 in
> 1-physical location, Node-4 in 2-physical location and ect. So if one
> physical location to fall out of the cluster, we get more breaks in ring
> than the theoretical minimum. I want to add an additional order in the
> process of constructing a ring so that ensure physically close nodes were
> close in topology.
>
> 2016-11-24 0:01 GMT+03:00 Denis Magda <dm...@apache.org>:
>
>> Here are some ticket you may want to start with
>> https://ignite.apache.org/community/contribute.html#pick-ticket
>>
>> —
>> Denis
>>
>> On Nov 23, 2016, at 5:33 AM, Dmitriy Setrakyan <ds...@apache.org>
>> wrote:
>>
>> Hi Alex,
>>
>> Welcome to the Ignite community!
>>
>> Once you are familiar with the contribution process, please see if there
>> is
>> a specific Jira ticket you would like work on. We will help you get
>> started.
>>
>> D.
>>
>> On Wed, Nov 23, 2016 at 2:24 AM, Vladislav Pyatkov <vpyatkov@gridgain.com
>> >
>> wrote:
>>
>> Please read https://ignite.apache.org/community/contribute.html#contribu
>> te
>>
>> Looking forward for your contributions!
>>
>> On Wed, Nov 23, 2016 at 1:15 PM, Александр Меньшиков <
>> sharplermc@gmail.com
>>
>>
>> wrote:
>>
>> I want to contribute.
>>
>> 2016-11-23 13:04 GMT+03:00 Vladislav Pyatkov <vp...@gridgain.com>:
>>
>> Hi Alex,
>>
>> This is Apache Ignite development community.
>> Are you want to contribute to Ignite or you are have any suggestion?
>>
>> On Wed, Nov 23, 2016 at 12:41 PM, Александр Меньшиков <
>> sharplermc@gmail.com>
>> wrote:
>>
>> Hello
>>
>>
>>
>>
>>
>>
>
>