It looks like the problem starts since the ecm handler patch 11.9 was implented.
1. Size of cache increased 4 times
2. CPU usage is more than doubled
3. Cacheex loop
It looks like the problem starts since the ecm handler patch 11.9 was implented.
1. Size of cache increased 4 times
2. CPU usage is more than doubled
3. Cacheex loop
This explains the unreasonable increase in cache size
All users have a seperate group. All readers get all groups except their own user-group. Put all users and readers in the same group as a test. Cache-size is the same as before but the all received cache is pushed as reported by the OP.Originally Posted by CapNCooK
@bunbunsat: please explain to me howto make 450 groups ?
Unfortunately it looks like the oscam developers don't care about the cacheex loop while it is so simple to solve the problem.. just get rid of the ecm handler patch in the later builds... Maybe someone can write a patch to make ecm handler patch undone ?
Patience is a bitch. If it is so simple, what holds you back to resolve it today ?
Get rid of the ecm_handler ? Just run 8894.
You have no idea how much time is beeing spend currently..
Didn't mean to be rude. But the thing is.. someone wrote a ticket about it and it was immediately closed. So I reopened the ticket and guess what it was immediately closed again as "works for me" ??. So I reopened it again. Now finally the ticket stays open. Since the ticket was closed again and again you already get a feeling that the problem was not taken seriously.
You are right about running 8894, I indeed reverted back to 8894. But more than half my clients still send back my own cache because they use later builds than 8894. So running 8894 only works if all your clients do the same. So till the problem is solved, there will be an extremely high bandwidth usage. Of course I appreciate the work of the developers but what I don't get is why an experimental patch that causes such a huge problem is not immediately taken out and provided as an optional patch like before till it is 100% sure that it is bug-free. For all those who are running cacheex on a VPS with limited traffic usage this is a huge problem. For me, I suddenly had 5TB traffic in 2 weeks.... Anyway.. I agree, all we can do is wait...
I test things for blueven, so speak to him quite a lot. Will mention the urgence to this in SVN, because i agree the overall bandwidth increase is a serious concern.
Furthermore categorically proven loopback, there and another nasty bug: Memory leak
It would be better dase looking fine solution to both problems.
Yes. Also known already.. Not sure this is directly related to ecmhandler.. You can confirm ?
Thx
Posting Permissions
Reply With Quote