optimal settings for 0963

[sky190176]

delete post

[sky190176]

delete post

[TV_LICENCE]

I believe oscam will go for the fastest ECM in your case..... so maybe your local is giving much faster ECM that from cache, that's why it's using a local instead of cache

[sky190176]

delete post

[CapNCooK]

I believe oscam will go for the fastest ECM in your case..... so maybe your local is giving much faster ECM that from cache, that's why it's using a local instead of cache

No, that's not correct.

Oscam will always check cache1/2/3 first, and then starts to pull from proxies.
The "Prefer Local Cards" setting does not apply to cache-ex.

If another client asks the ECM, before it has arrived in cache, it cannot answer from cache. (sounds logical eh )
I think the latency of the new connected client is lower, and therefore oscam got the request earlier than your other client.

You can try to fix this by modifying this parameter in oscam.conf:

cacheexwaittime = 50 (default)

By this setting, Oscam will WAIT [50 ms] to receive a cache1/2/3-entry for the request. If not received within [50ms] than the request will forward to a proxy reader.

Don't set this too high, e.g., if you set it to 350, the request will:

1) ECM Request => latency network 20ms
2) Oscam receives requests and waits for cache 350ms (total latency: 370ms)
3) No cache = > start to read from proxy/card => (total latency: 380 ms)
4) Ecm request received from card in 250ms (total latency: 630ms)

and finally:

5)Answer sent to client => latency network 20ms (total latency: 650ms)

[sky190176]

delete post

[CapNCooK]

No..

cacheexwaittime = as explained..

The "cachedelay" parameter is just a delay in returning a found ECM from cache.
Having cachedelay=0, your cache will respond in 5ms, 2ms, 6ms etc..

This seems good, but actually is not, because the loadbalancer on the demanding client's oscam will save these low ecm answers, preferring that reader for all requests.. (having loadbalancermode=1)..

So it's good to put a little delay (=50, =75 or so)... So the loadbalancer can save these results as avarage..

Hope this helps!