Svr1 , Sv2 , Sv3 , are all one unique STORE of your data. this is the
main goals of memcached, all data is distributed in all servers, and you
dont know where is your particular key:data, but, you know when you need,
you can get fast your data. if you need to grow or scale, you can add
servers as many as you like. the CACHE, will have a SUM of all individual
RAM in each server.
Remember, this is CACHE, not a DATABASE. so you need to have some
considerations. (**)
(**)you need to consider this: In this scenario, if you turns off a Svr1, ,
you lost access to all keys in that particular Server. (all are lost).
that is not a problem, because your original application, has used
memcached successfully. so memcached will do the same think with you new
servers.
In many scenarios, CACHE means disposable. if this is not your case so you
need to make some research.
(**)CACHED, information is stored for a limited timeframe, is if its needed
again, you need to recalculate again. (this is very useful).
Replication means, have multiple instances of memcache with same data. BUT
ONLY ONE IS ACTIVE the others are PASSIVE.
but i think in a CACHE solution, only in some particular cases, replication
can be useful, in the majority, the standard way (1 memcache cluster) is
enough.
On Friday, October 12, 2012 11:16:57 AM UTC-5, Kiran Kumar wrote:
>
> Thank you very much for your efforts and the great explanation .
>
> *I have two questions with respect to your answer ?? Please let me know *
>
> 1. The different memcache servers Svr1 , Svr2 and Svr3 will be as
> active/active clustering or active/passive clustering , and what is the
> default behavior related to xmemcahed framework?
> 2. What exactly does Replication meant in your answer ?? ( Will these
> three servers will not have same copy of Data automatically ) , i was
> asking this because if is shutdown servers say Svr1 and Svr2 ,and still i
> am able to get the data from Svr3 . Is this different from what replication
> means ??
>
>
> On Friday, 12 October 2012 21:22:06 UTC+5:30, ArtPort wrote:
>>
>> Kiran Kumar, you can do what you intended to do, but not in the way you
>> intended to do.
>> memcached has already a useful way to support scalability, with increment
>> of performance lineally.
>>
>> if you already have "one Memcached for the entire application".
>> and work well, so, if you need scale your application, only
>> you need to do is add some servers to the previous memcached cluster.
>>
>> As You Stated, This Configuration is Working OK.
>>
>> Streaming Application -----Write-----> Memcached
>> ------Read/Delete----> ClientReads
>> |
>> |
>> |
>> Svr1
>>
>>
>> Now, You have grow, and need make a scale of your Memcached.
>>
>> The easiest way is:
>>
>> Streaming Application -----Write-----> Memcached
>> ------Read/Delete----> Client
>> / | \
>> / | \
>> / | \
>> Svr1 Svr2 Svr3
>>
>> You can do this, with this instructions:
>> *//PHP
>> $MEMCACHE_SERVERS = array(
>> "10.1.1.1", //Svr1
>> "10.1.1.2", //Svr2
>> "10.1.1.3", //Svr3
>> );
>> $memcache = new Memcache();
>> foreach($MEMCACHE_SERVERS as $server){
>> $memcache->addServer ( $server );
>> }
>> *
>> Code taken from:
>> Here<http://stackoverflow.com/questions/4717559/multiple-memcached-servers-question>
>>
>> This way, the performance is incremented because is using multiple
>> servers to do the task (writes/reads/deletes),
>> So, the hole capacity, (if each Server has 8Gig Ram, so the Memcached
>> cluster have 24Gig RAM) and boosted in performance.
>> Again, you do need to do what you initially intended to do. all is
>> already done with memcache in the current version.
>> If you need Replication, that is another problem. but you do not mention
>> that.
>>
>> Do not use the same web server as memcache server, use as separated
>> servers exclusively for memcached as the graph
>> shows. (may be you think differently).
>>
>> On Friday, October 12, 2012 9:17:25 AM UTC-5, Kiran Kumar wrote:
>>>
>>> Initially there is one Memcache for the entire application , now we
>>> expect some 50000 users on to our application , so as part of re modeling
>>> we are going with 2 Memcache servers , which will be working independently
>>> to suppourt scalibility , and as the exisiting application is designed in
>>> such a way that one part of the Application will be writing to both the
>>> instances of the application .
>>>
>>> On Thursday, 11 October 2012 20:28:06 UTC+5:30, Kiran Kumar wrote:
>>>>
>>>> I am working on a heavy traffic web site , where there will be GB's of
>>>> data written per minute into our Memcache . So we have decided to use two
>>>> separate instances of Memcache for the application .
>>>>
>>>> Right now the setup is that , there is NO clustering between Memcache1
>>>> and Memcache2 because , Memcache1 caches 50% of the data and Memcache2
>>>> caches the other 50% of the data.
>>>>
>>>> Memcache1 Memcache2
>>>> \ /
>>>> \ /
>>>> \ /
>>>> \ /
>>>> \ /
>>>> \ /
>>>> CustomerData
>>>>
>>>> So right now as per the set up , there are two Memcache instances for a
>>>> single application .
>>>>
>>>> Now my question is , once we recive a value inside the application ,
>>>> which writes/sets to both the Memcache instances , assume that if a key is
>>>> read one of the instance of Memcache - 1 , i need to delete the same key
>>>> on
>>>> the other instance of memcahce also at the same time , so taht they will
>>>> be
>>>> in sync with each other .
>>>>
>>>> As per the code point of view once a value is read from Memcache , i am
>>>> deleting that key .
>>>>
>>>
