How to use Kademlia demo?

[maidsafe-archive]

How to use Kademlia demo?

[dirvine]

We have rewritten much of the kademlia codebase now. I am not sure if kaddemo has been done yet. I will ask the guys on Monday. It looks liek the problem you are seeing should not happen and is not your fault (i.e. you are doing the right thing). There is a couple of base classes with derived classes for additional functionality such as a secured kadmelia network. In that case you would need to implement a 'securifier' however a default null securifier should be in place I believe. As I said I will check that out on Monday.

[deandingcn]

Thanks guys.I have tried maidsafe-dht v.28 with maidsafe-common v6,and after compiled,I found there's no kademliademo such this file in the folder(Only TESTkademlia and TESTtransport).
So I used TESTkademlia instead, and after running a little while,it reported a segment error.:(
By the way,what does it exactly mean by " Global test environment set-up" which was displayed on the screen after I runned maidsafe-dht?

[dirvine]

That's part of the google test setup. It allows you to set up a complex environment (or one that takes a while) and reuse it for many tests. We use this a lot and especially in kad code as setting up a network on a machine is a pain to say the least. This script does that for us. If you look at the test itself you will see the environment being set up.
Do you have any info on the segfault at all?

[deandingcn]

when everytime I runned testkademlia here,segfault will occur.
[----------] 12 tests from DataStoreTest
[ RUN ] DataStoreTest.BEH_KAD_StoreUnderEmptyKey
segment error

So "Global test environment set-up" is nothing to do with Kademlia network environment,right?

[dirvine]

Thanks for this we will try and recreate this. If post can you post the compiler you are using.

the setup phase does set up a kademlia network in some tests. I need to look at the code but am out of the office just now.

[Fraser999]

Hi there,

I've run this test a few times and have been unable to reproduce the segfault. Could you let us know the platform and compiler you're using please?

Thanks,

Fraser.

[deandingcn]

Hi guys,
The compiling and testing environment is as below
/*********************************************************
GCC 4.4.5;
G++ 4.4.5;
Kernal: 2.6.32-5-686(Debian 2.6.32-34squeezel)
*********************************************************/
And again,Could anybody tell me how to use KademliaDemo?
Thanks,
-Deane.

[Fraser999]

Hi Deane,

I'm afraid we can't reproduce your error here. We have a few Linux boxes that have run the test successfully, but we're all using GCC 4.5 or 4.6. However, I'm not sure that would explain the segfault you're seeing.

Can you possibly re-run the test using GDB and let us see the output of "where" once it crashes?

As for KademliaDemo, we haven't got it working yet. As David mentioned we underwent a fairly significant code rewrite recently, and KadDemo hasn't made it to the top of the list yet. We have a couple of devs working on related tasks here just now, so it shouldn't be too long hopefully until we have KadDemo back up and running. If you like I can let you know when we're there?

Thanks,

Fraser.

[deandingcn]

Hi Fraser,
Good morning,
I runned TESTkademlia_d using GDB again,
Below is the GDB information,I only took the faliure section(actually when it reported failure,I stopped the running)
/_GDB Test_*******
[ RUN ] DataStoreTest.BEH_KAD_StoreUnderEmptyKey
unknown file: Failure
C++ exception with description "Integer: missing Max argument" thrown in the test body.
[ FAILED ] DataStoreTest.BEH_KAD_StoreUnderEmptyKey (3362 ms)

**********************************************************************************************************/
I wish it could make some help.

-Deane

[deandingcn]

Sorry,the last exception did not exactly cause segfault, I made a mistake.
Below is the right place.
/********************************************************************************************************
[Thread 0xb32ffb70 (LWP 2914) exited]
[Thread 0xb2afeb70 (LWP 2915) exited]
[Thread 0xb22fdb70 (LWP 2917) exited]
[Thread 0xb1afcb70 (LWP 2918) exited]
[Thread 0xb12fbb70 (LWP 2919) exited]
[ OK ] ServicesTest.BEH_KAD_MultipleThreads (9375 ms)
[ RUN ] ServicesTest.BEH_KAD_SignalConnection
AAAAAAAaa
[New Thread 0xb12fbb70 (LWP 2920)]
[Thread 0xb12fbb70 (LWP 2920) exited]
[New Thread 0xb12fbb70 (LWP 2921)]
E0629 10:17:20.634300 2921 crypto.cc:212] Crypto::AsymCheckSig - BER decode error
W0629 10:17:20.634367 2921 service.cc:485] Failed to validate request_signature
[Thread 0xb12fbb70 (LWP 2921) exited]

Program received signal SIGSEGV, Segmentation fault.
0x08a3fdf1 in ~service_registry (this=0xb3307f88,
__in_chrg=)
at /home/dingyao/boost_1_46_1/boost/asio/detail/impl/service_registry.ipp:42
42 service->shutdown_service();
(gdb)
**************************************************************************************************/
Sorry for the fuss,:)
Deane.

[deandingcn]

Hi,Guys,
I got another question.
Say,could maidsafe-dht still work if nodes are behind NAT and the NAT gateway device happens not to support UPnp or NAT-PMP?

[dirvine]

No prob

we also use hole punching for udp packets to traverse a router. Yet unimplemented fully as is the ability for nodes to listen also on tcp 80 and 443 to allow relay capability. This is all work in progress, but to answer the question yes it will do, but not quite yet.

Hope that helps
David