Opened 9 years ago

Closed 9 years ago

#4094 closed defect (fixed)

can't connect to cccam cards at hop bigger than 1

Reported by: Samuel Owned by: Samuel
Priority: critical Component: Loadbalancer
Severity: high Keywords: service search server cccam
Cc: Sensitive: no

Description

Revision

10357

Issue Description

proxy can't connect to cccam cards at hop>1

When the issue occurs

Using a CCcam server configured in oscam with hop 1 and 2 cards reshares.
Proxy only connects to CCcam cards at hop1.

How the issue is reproducable

---oscam.conf
reshare_mode=3, minimize_cards=2 and keep_connected = 1

---oscam.server
cccmaxhops = 5
cccreshare = 5
ccchop = 5
cacheex_maxhop = 5

---oscam.user
cacheex_maxhop = 5
cccmaxhops = 5
cccreshare = 5

Only connect to CCcam cards at hop1 from configured CCcam server.

Attachments (4)

oscam.log (75.5 KB ) - added by Samuel 9 years ago.
Oscam log file with only hop1 cards being used
oscam-r9948.log (87.7 KB ) - added by Samuel 9 years ago.
cards hop>1 not being used
oscam-r9972.log (88.2 KB ) - added by Samuel 9 years ago.
cards hop>1 not being used
oscam-r10412.log (97.4 KB ) - added by Samuel 9 years ago.
cards hop>1 not being used

Download all attachments as: .zip

Change History (18)

by Samuel, 9 years ago

Attachment: oscam.log added

Oscam log file with only hop1 cards being used

comment:1 by Samuel, 9 years ago

Summary: caid:srvid search blockedcan't connect to cccam cards at hop bigger than 1

comment:2 by Samuel, 9 years ago

Also present in log attached, there is another problems registered.

With same conditions pointed above as reshare_mode=3, minimize_cards=2 and keep_connected = 1

I)
When a card is selected after first iterations, when a srvid fails in that card it just stays blocked at state not found, without trying to find it in other active servers.
I already define each cccam server in different group and then associated the client with all groups.

II)
When all CCcam servers defined at oscam.server belong to the same group I realised that services=!<service> were being somehow merged and so when server in proxy was changed the blocked services continues to be blocked in the other card. Setting each server with different group number solved this issue.

in reply to:  2 comment:3 by AbrahaM, 9 years ago

Replying to samuelbernardo:

Also present in log attached, there is another problems registered.


I'm suspecting something. Please test following oscam builds:

r9948
r9972

and please report back results. And DO NOT user minimizecards=2, perform tests with minimizecards=0 and =1

Last edited 9 years ago by AbrahaM (previous) (diff)

comment:4 by Aeon, 9 years ago

the sid merging has been adjusted, please re-test

http://www.streamboard.tv/oscam/timeline

comment:5 by Samuel, 9 years ago

hello,

I'm sorry the delay of my feedback.
Today I'll give you the output of the tests you requested.

thks

comment:6 by Samuel, 9 years ago

With OSCam r10412 the sid problem could not be reproduced as before.
I'm testing over last days with this version and seems to be stable.

About cards hop issue, in webif status it appears the following information for a CCcam server (proxy) with cards at hops 1 to 4:
CONNECTED
(12 of 68 cards)

where 12 are cards at hop1 and the remaining are hop > 1.
Since the defect I anounce remains in OSCam r10412, I'll test with AbrahaM instructions to see if there is any difference and post here the logs.
I'll check also the webif code to understand how connection behavior is implemented.

Last edited 9 years ago by Samuel (previous) (diff)

by Samuel, 9 years ago

Attachment: oscam-r9948.log added

cards hop>1 not being used

by Samuel, 9 years ago

Attachment: oscam-r9972.log added

cards hop>1 not being used

by Samuel, 9 years ago

Attachment: oscam-r10412.log added

cards hop>1 not being used

comment:7 by Samuel, 9 years ago

I just added the log files for 3 versions of oscam.
In all of them, only cards at hop 1 are being connected.
Is this the expected behavior?

If I force to only use cards at hop > 1 the proxies can't connect with any card.

comment:8 by Poedel, 9 years ago

'minimizecards=2' gives the issue here. With this setting on 0 or 1, no problem.

comment:9 by Samuel, 9 years ago

In my case, even with minimizecards=0 or 1 it continues to appear in proxy status:
CONNECTED
(12 of 68 cards)

That corresponds to only hop1 cards being connected and not the others.
poedel can you give me some more details about your test with minimizecards=0 or 1?

comment:10 by Poedel, 9 years ago

I think you read the proxy reader status wrong, I don't exactly know why OSCam says 12 of 68 cards, but it is connected, sees 68 cards and it can all use it. Zap to a channel which will open only with a hop 2 card, and you will see this works.

in reply to:  10 comment:11 by Samuel, 9 years ago

Replying to poedel:

I think you read the proxy reader status wrong, I don't exactly know why OSCam says 12 of 68 cards, but it is connected, sees 68 cards and it can all use it. Zap to a channel which will open only with a hop 2 card, and you will see this works.

Yes, that's right. I actually review that currenthops now appear with value > 1, when using minimizecards=0. But I don't know if it's another bug in webif, since it displays that only hop 1 cards are connected and it seems not to be the actual state.

With minimizecards=1 it chooses only cards with lesser hops so that is not the best choice to load balance all cards.

With minimizecards=2 as I understood, it aggregates all cards from same sids, so I thought it would be the best in terms of efficiency. So it's the current behavior expected, when it discards all cards with hop > 1 ?

comment:12 by stefansat sat, 9 years ago

with minimizecards=2 it agreates all prid to one for a specific caid.. Result tremendous amount off fake ecm's are sended.

means by caid 100 every ecm will be sended to all avbl prid.

so ecm for card 0100:006a will be sended to 006b 006d and all others provid's with caid 0100. same for caid 0500

in reply to:  12 comment:13 by Samuel, 9 years ago

Replying to stefansat:

with minimizecards=2 it agreates all prid to one for a specific caid.. Result tremendous amount off fake ecm's are sended.

means by caid 100 every ecm will be sended to all avbl prid.

so ecm for card 0100:006a will be sended to 006b 006d and all others provid's with caid 0100. same for caid 0500

Thanks for the clarification. I understood cards aggregation wrongly.
So for load balancing minimizecards=o will be the best choice. Afterwards lb will do the work selecting the cards with lesser load.

comment:14 by Aeon, 9 years ago

Resolution: fixed
Status: newclosed
Note: See TracTickets for help on using tickets.