Client Vs Server Configuration

[Victor <vi...@gmail.com>]

Hi,

I am planning to use oob server nodes (i.e. started via ignite.sh) pass a
configuration xml containing the basic discovery ip only.
In my client code, i plan to define the Cache configuration, query fields
and additionally my custom RDBMS support via implementing the
CacheLoader->load(...) method for cache miss cases.

Given this setup/configuration, I have tried the Cache/Query field
configuration part and that works, i have to yet try out the RDBMS bit via
CacheLoader. Had a couple of basic queries,

1. Is this configuration model doable and will it seamlessly work?
2. Are there any pros/con's with not having the config on the sever end,
performance, n/w hops, etc?

Thanks,
Victor



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Client Vs Server Configuration

[Ilya Kasnacheev <il...@gmail.com>]

Hello!

I don't see why you can't use Data Streamer here. With allowOverwrite=false
you don't even need 'get' checks.

CacheLoader is pulling data from underlying database, not sure it is the
right choice for push-style data loading.

Regards,
-- 
Ilya Kasnacheev


вт, 21 янв. 2020 г. в 20:50, Victor <vi...@gmail.com>:

> If i understand correctly, the usecase for data streamer would be to pre
> load
> large/bulk data into cache.
>
> The usecase that i am looking at is, say we have limited cache, where say
> 50% data is pre-loaded and here the data stream will come into play and
> remaining 50% is in DB. The "get" call's that result in a cache miss, will
> load the data from the DB and populate it in the cache for subsequent
> lookups.
>
> My plan was to use CacheLoader for this Cache miss case. I don't think Data
> Streamer would be a good choice here, would it?
>
> And if CacheLoader is the only way to do this then i will have to implement
> the server nodes as well, right?
>
> Regards,
> Victor
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/
>

Re: Client Vs Server Configuration

[Victor <vi...@gmail.com>]

If i understand correctly, the usecase for data streamer would be to pre load
large/bulk data into cache.

The usecase that i am looking at is, say we have limited cache, where say
50% data is pre-loaded and here the data stream will come into play and
remaining 50% is in DB. The "get" call's that result in a cache miss, will
load the data from the DB and populate it in the cache for subsequent
lookups.

My plan was to use CacheLoader for this Cache miss case. I don't think Data
Streamer would be a good choice here, would it?

And if CacheLoader is the only way to do this then i will have to implement
the server nodes as well, right?

Regards,
Victor



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Client Vs Server Configuration

[Ilya Kasnacheev <il...@gmail.com>]

Hello!

I think that CacheLoader will be executed on server nodes. If you need data
loading to happen on client node, better use Data Streamer.

Otherwise, sounds OK, location of cache configurations is not important
once caches are started.

Regards,
-- 
Ilya Kasnacheev


вт, 21 янв. 2020 г. в 02:39, Victor <vi...@gmail.com>:

> Hi,
>
> I am planning to use oob server nodes (i.e. started via ignite.sh) pass a
> configuration xml containing the basic discovery ip only.
> In my client code, i plan to define the Cache configuration, query fields
> and additionally my custom RDBMS support via implementing the
> CacheLoader->load(...) method for cache miss cases.
>
> Given this setup/configuration, I have tried the Cache/Query field
> configuration part and that works, i have to yet try out the RDBMS bit via
> CacheLoader. Had a couple of basic queries,
>
> 1. Is this configuration model doable and will it seamlessly work?
> 2. Are there any pros/con's with not having the config on the sever end,
> performance, n/w hops, etc?
>
> Thanks,
> Victor
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/
>