Memcached-backed cache location?

[valentijnvenus]

This is a FEATURE request:

Instead of using Nuster's cache (in-memory or disk persistent solutions), would you have any pointers if Memcached is to be used?

e.g.:
nuster cache on data-size 10m memcached address:11211

We might be able to contribute.

[jiangwenyuan]

Hi, there's no plan to add support for memcached backed persistence, but the contribution is welcomed:)

[AndreiG6]

If anything redis would be better considering it can store in both ram and disk

[valentijnvenus]

Is redis compatible with the memcache-protocol? If so, one could use any stack as long as it is compliant, e.g. incl. couchbase.

@jiangwenyuan any pointers to the c-code you'd suggest studying in more detail?

[jiangwenyuan]

@valentijnvenus Just out of curiosity, why do you need memcached-backed persistence?

Here's some code you might need to look into

https://github.com/jiangwenyuan/nuster/blob/master/include/nuster/nuster.h#L57 for parse nuster in global section

https://github.com/jiangwenyuan/nuster/blob/master/src/nuster/parser.c#L880 for parse nuster rule

https://github.com/jiangwenyuan/nuster/blob/master/src/nuster/cache/filter.c#L116 the entry of request/response, checking key existence

https://github.com/jiangwenyuan/nuster/blob/master/src/nuster/cache/filter.c#L290 caching data

etc

[packeteer]

@valentijnvenus Just out of curiosity, why do you need memcached-backed persistence?

It would be good to have the ability to use Nuster in a cluster setup ie. replicate data to other nodes.
In my use case, I split the traffic 3 ways (for redundancy and load balancing) before it even hits Nuster, which dilutes the effectiveness of any caching

[valentijnvenus]

Indeed.

…

On Tue, 24 Dec 2019, 00:40 packeteer, ***@***.***> wrote: @valentijnvenus <https://github.com/valentijnvenus> Just out of curiosity, why do you need memcached-backed persistence? It would be good to have the ability to use Nuster in a cluster setup. In my use case, I split the traffic 3 ways (for redundancy and load balancing) before it even hits Nuster, which dilutes the effectiveness of any caching — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#46?email_source=notifications&email_token=ABCKJDJRUTB77RQUTW7FYTTQ2FEABA5CNFSM4IU3IAPKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHSEC6I#issuecomment-568607097>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABCKJDJKLRBPILPXIQVQ36LQ2FEABANCNFSM4IU3IAPA> .

[packeteer]

fwiw, I think something like ZeroMQ or NNG might be a better fit for this
https://zeromq.org/
https://nng.nanomsg.org/

[jiangwenyuan]

iiuc, ZeroMQ is a message queue, how can we find data by key in such system?

BTW, I'm working on using redis as 3rd backend store.

[packeteer]

oh ok, I was thinking the service would only be used to transfer content between running processes, not used as a backend.
maybe my misunderstanding of Nuster internals?

[jiangwenyuan]

I see, do you mean use that as a sync mechanism?

[packeteer]

yes, exactly

[jiangwenyuan]

Interesting, thanks for the info:) I would evaluate that along with haproxy peers an other options.

[igorescobar]

Redis would be a better start in my opinion.

Interest worldwide - last 5 years - Memcached (blue), Redis (Red):

So if you want to start somewhere I would bet on the right horse 😄