Cache write-through for backup entries

[Shaun Mcginnity <sh...@gmail.com>]

Hi,

I would like to implement cache write through for both primary and backup
entries. The use-case is to maintain a backup of the cache on disk on each
DDM node.

I have implemented CacheWriter but it is only called for primary entries,
not backup entries.

Is there a way that I can implement this other that periodically polling
all entries in the cache?

Regards,

Shaun

Re: Cache write-through for backup entries

[Shaun Mcginnity <sh...@gmail.com>]

Hi Val,

thanks, I will give that a try.

Regards,

Shaun

On Mon, Apr 4, 2016 at 7:31 PM, vkulichenko <va...@gmail.com>
wrote:

> Hi Shaun,
>
> If the data is small, you can simply switch to replicated cache. Will this
> work for you?
>
> -Val
>
>
>
> --
> View this message in context:
> http://apache-ignite-users.70518.x6.nabble.com/Cache-write-through-for-backup-entries-tp3861p3911.html
> Sent from the Apache Ignite Users mailing list archive at Nabble.com.
>

Re: Cache write-through for backup entries

[vkulichenko <va...@gmail.com>]

Hi Shaun,

If the data is small, you can simply switch to replicated cache. Will this
work for you?

-Val



--
View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Cache-write-through-for-backup-entries-tp3861p3911.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Re: Cache write-through for backup entries

[Shaun Mcginnity <sh...@gmail.com>]

Hi Val,

thanks for the reply.

My use-case is as follows: I have a small cache (say 100000 entries) across
3 processes. I don't have a shared backend store. If I have one primary and
one backup then I can recover from a failure of 1 processes. But I can't
recover from a failure of 2 processes. If I could write the cache to disk
then I could recover from 2 or more processes failing, although at the risk
of losing some data that was in-flight at the time of a crash.

I know it is an extreme case but it is being raised in our design
discussions.

Regards,

Shaun

On Fri, Apr 1, 2016 at 10:00 PM, vkulichenko <va...@gmail.com>
wrote:

> Hi,
>
> Store is not updated on backup nodes because Ignite assumes that the store
> is some kind of shared storage. Can you provide more details on your use
> case? How are going to use the data written to the store afterwards?
>
> -Val
>
>
>
> --
> View this message in context:
> http://apache-ignite-users.70518.x6.nabble.com/Cache-write-through-for-backup-entries-tp3861p3872.html
> Sent from the Apache Ignite Users mailing list archive at Nabble.com.
>

Re: Cache write-through for backup entries

[vkulichenko <va...@gmail.com>]

Hi,

Store is not updated on backup nodes because Ignite assumes that the store
is some kind of shared storage. Can you provide more details on your use
case? How are going to use the data written to the store afterwards?

-Val



--
View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Cache-write-through-for-backup-entries-tp3861p3872.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.