Help require with grouping the cacheex=3

[sunlight786]

ok.

so if i understand CapNcook then it will be as

local card( group=1 )

normal newcamd reader1 ( group=2 )
normal newcamd reader2 ( group=2 )
normal newcamd reader3 ( group=2 )
normal newcamd reader4 ( group=2 ) #every newcamd reader will be group =2


normal CCcam reader1 ( group=2 )
normal CCcam reader2 ( group=2 )
normal CCcam reader3 ( group=2 )
normal CCcam reader4 ( group=2 ) #every CCcam reader will be group =2
########################################

#####normal oscam.user will be as

[account]
user = normal-user1
pwd = pass1
group = 1,2,10 # here we include our local+newcamd+CCcam+all cach user group( which is 10 )

[account]
user = normal-user2
pwd = pass2
group = 1,2,10 # here we include our local+newcamd+CCcam+all cach user group( which is 10 )

[account]
user = normal-user3
pwd = pass3
group = 1,2,10 # here we include our local+newcamd+CCcam+all cach user group( which is 10 )

######################################## ########## ###########

###Now we are making cach-server##

[reader]
label = cachex=peer1
group = 1,2 ## do i need to add here group=10 ( this group is for all oscam cach user,s )

[reader]
label = cachex=peer2
group = 1,2 ## do i need to add here group=10 ( this group is for all oscam cach user,s )

[reader]
label = cachex=peer3
group = 1,2 ## do i need to add here group=10 ( this group is for all oscam cach user,s )

##################################

####Now we are in cach user ### All cach user will be Group = 10

[account]
user = cach-peer1
pwd = pass-1
group = 10

[account]
user = cach-peer2
pwd = pass-2
group = 10


[account]
user = cach-peer3
pwd = pass-3
group = 10

######################################## ########## ########

Please Chech my Groups Now and tell me if I understand right?

[CapNCooK]

Yeah, well without the whole example thing, its rather simple.

It doesnt matter how you set your groups for normal sharing.

For cache:

Put _one_ group for all incoming cache. So on each cache-user (in mode 3) put eg. group 60.
If you receive cache in mode-2, put 60 on the reader.
Also put 60 on each user that you entitle to use the cache.

Then for outgoing cache, set all (non-cache) groups on each outgoing cache user/reader, and include group 60 _if_ you want to forward incoming cache to your cache-peers.

Looking at your example, you used the rules above, so yes, its correct.

[turbopower]

I think if we put all peers and readers in one group will be the same. Loopback is in the past from couple of months ago so there is no make sense to complicate our configurations.

[CapNCooK]

I think if we put all peers and readers in one group will be the same. Loopback is in the past from couple of months ago so there is no make sense to complicate our configurations.

It has never existed.. Its widely spreaded non-sense

However, if you assign multiple groups to an incoming cache user, you will duplicate up load. (since the cache is registered in each group, so duplicate for each added group).

All will go up.. CPU load, outgoing traffic. etc.etc.. De-Duplication won't work.. not good..

one group is key.. (unless functionally needed to route cache)

[turbopower]

It has never existed.. Its widely spreaded non-sense

However, if you assign multiple groups to an incoming cache user, you will duplicate up load. (since the cache is registered in each group, so duplicate for each added group).

All will go up.. CPU load, outgoing traffic. etc.etc.. De-Duplication won't work.. not good..

one group is key.. (unless functionally needed to route cache)

Yes that is what I mean. In this way of thinking there is not reason to exclude some members from cache. Cacheex keep our locals from overheat and this is his big advantage.

About the groups in my previous config I limited every account in separate unique group and exclude this group in the account belong reader. This make my config very strict but imagine if you have more than 50 exchanges.

[sunlight786]

I have tried 2 diffrent group setup from this thread atleast 4 hours with same amount cards.

Test 1 = post 3 of this thread.
Test 2 = post 11 of this thread

here is the result so far I have seen ( remeber my entension is how can I get better relyable flow of hit from cach I am getting from all arround )

on oscam webinfo cacheex page this is what i found

Test 1=same amount of cach I got, similer hit I got, 0 errer,producing too much CW diffrent , cach size showing big example 1325

Test2=same amount of cach I got, similer hit I got,0 errer,very little CW diffrent(alomost none) cach size showing less example 425

##seems Test 2 set up is better and true. explain plz if I am wrong

in this case more than 64 peers can be used.

[CapNCooK]

I have tried 2 diffrent group setup from this thread atleast 4 hours with same amount cards.

Test 1 = post 3 of this thread.
Test 2 = post 11 of this thread

here is the result so far I have seen ( remeber my entension is how can I get better relyable flow of hit from cach I am getting from all arround )

on oscam webinfo cacheex page this is what i found

Test 1 = similer amount of cach I got, similer hit I got, 0 errer,producing too much CW diffrent .
Test 2 = similer amount of cach I got, similer hit I got, 0 errer,very very little CW diffrent ( alomost none )

in this case more than 64 peers can be used.

If you want more hits, increase your waittime to the edge per caid. (increase upto freeze)
Also increase cacheex_maxhops to increase cachesize.

You dont have to test anything, its not rocket science.
Just set it as explained. (use one group for incoming cache, anything else is faulty).

CW Diff doesnt have to be a problem, everyone got a few of these.

[sunlight786]

got your point. I am keeping maxhops minimum due to the Bandwith .