aMule Forum

English => en_Bugs => Topic started by: vdb on January 03, 2006, 09:15:20 PM

Title: Connection to Kad lost
Post by: vdb on January 03, 2006, 09:15:20 PM: Developers, first off congratulations on this release.

However, it seems that the Kad disconnect bug is not entirely squashed yet, it seems the daemon loses connection after a while. This time I am actually downloading more then 5 files, most with more then 100 sources:
[me@mine .aMule]$ cat logfile | grep Kad
2006-01-03 08:03:39: Read 0 Kad contacts
2006-01-03 08:03:40: Kad started.
2006-01-03 08:03:41: Kad stopped.
2006-01-03 08:03:41: Connected to Kad (firewalled)
2006-01-03 08:03:57: Connected to Kad (ok)

2006-01-03 18:03:41: Connected to Kad (firewalled)
2006-01-03 18:03:42: Connected to Kad (ok)
2006-01-03 19:02:34: Disconnected from Kad
2006-01-03 21:29:14: Read 145 Kad contacts
2006-01-03 21:29:14: Kad started.
2006-01-03 21:29:35: Kad stopped.
2006-01-03 21:29:41: Read 93 Kad contacts
2006-01-03 21:29:41: Kad started.

The last 2 entries where when I tried to start Kad from the amulecmd command line with the sequence "connect kad", "disconnect kad" and "connect kad" again. It seems the only way to gain Kad connectivity again is stopping and starting the daemon.

Any suggestions?
Title: Re: Connection to Kad lost
Post by: Bourdieu on January 03, 2006, 09:34:32 PM: I have the same behavior here with aMule 2.1.0 using amuled on linux slackware 10.2.

I can connect to kad immediately without any problem, but after approximately 4 hours I was disconnected and couldn't reconnect anymore.

Hope it'll help to find the solution.
Title: Re: Connection to Kad lost
Post by: Radek on January 04, 2006, 12:20:01 PM: Same at my place.

But it is getting better (or so it seems). This time I lost connection after 43 hours. All the tests with CVS before this release lasted only for max 20 hours, as far as I can remember.

I am using amuled 2.1.0 on Debian Sarge
Title: Re: Connection to Kad lost
Post by: papaf on January 04, 2006, 03:17:14 PM: Hi all,
I just recently switched from using amuled (up to 2.1.0) to using the monolithic app.
While I was having the same problems, with kad often offline, I have to say that without
changing anything else, the monolith app seems to be ok. I never got disconnected from
kad, and it's been nearly a week online now.

Also, I have to report that search operations work much better than the remote gui one.
Even non-kad searches.

Some of you that have those kad problems, can you verify if you still have the problems
without using amuled? Just out of curiosity...

Bye, Fabio.
Title: Re: Connection to Kad lost
Post by: vdb on January 05, 2006, 12:33:41 PM: Yep, still having the problem with kad disconnect. I have lowered my max connections, and connections/second, but it seems to make no difference:
[me@mine .aMule]$ cat logfile | grep Kad
2006-01-03 21:39:09: Read 33 Kad contacts
2006-01-03 21:39:10: Kad started.
2006-01-03 21:39:10: Kad stopped.
2006-01-03 21:39:10: Connected to Kad (firewalled)
2006-01-03 21:39:14: Connected to Kad (ok)
2006-01-03 23:31:13: Disconnected from Kad

Seems that this time it already disconnected after merely 2 hours of operation, where the daemon crashed after just over 1.5 days of operation:
[me@mine .aMule]$ cat logfile
2006-01-03 21:39:08: AICH Thread: Syncronization thread started.
2006-01-03 21:39:08: Connecting
2006-01-03 21:39:08: Servers: Trying to connect

2006-01-05 09:36:11: Received 4 new servers
2006-01-05 09:36:11: Saving of server-list completed.

I can't really try the monolith, since the machine I'm running it on doesn't have a monitor nor X :)
Title: Re: Connection to Kad lost
Post by: IWantMyMule on January 05, 2006, 01:14:24 PM: I've got the same problem with the monolithic mule.
Title: Hi all
Post by: acehigh on January 08, 2006, 04:31:56 PM: I want to suggest all the people who found this problem the reading of these 2 topics:

thread (http://forum.amule.org/thread.php?threadid=8361&sid=)

and

thread (http://forum.amule.org/thread.php?threadid=7841)

I'm presently running the CVS 06012006 but the problem still remains.

I'm suggesting again to the developers to put something significant in the logfile when amuled decides to disconnect itself, so we users can put the log in the forum and help the developers.
Title: Re: Connection to Kad lost
Post by: inhumanity on January 09, 2006, 02:13:47 AM: I have the same problem using amuled+amuleweb kad will disconnect at seemingly random intervals and not auto reconnect nor is it possible to manually reconnect it using amulecmd only a restart of amuled gets it running again.
Title: Re: Connection to Kad lost
Post by: DigitoErgoSum on January 11, 2006, 06:50:03 PM: Same problem for me using amuled+amuleweb. I'm using amuleCVS (as always 8) )
Title: Re: Connection to Kad lost
Post by: kreegee on January 12, 2006, 06:19:36 PM: same here with amuled and 2.1.0
Title: Re: Connection to Kad lost
Post by: PxPert on January 13, 2006, 08:25:48 PM: Same problem here... ;(
Title: Re: Connection to Kad lost
Post by: Kry on January 13, 2006, 09:24:39 PM: As I said before, it doesn't happen here. I can have Kad enabled and run for days without a single disconnection. So I can't do anything on my side. It will ahve to be another dev. But I' the Kad dev, so...
Title: Re: Connection to Kad lost
Post by: vdb on January 18, 2006, 11:03:29 AM: Is there any way of improving the Kad connect function for the daemon? It seems that ordering it to disconnect/connect works nicely if it actually is connected to Kad, but it does not work if it has automatically disconnected itself, see previous post.

Maybe something to look at in the upcoming CVS'es?
Title: Re: Connection to Kad lost
Post by: PxPert on January 18, 2006, 01:53:23 PM: Very strange think... Now kad works perfectly. The only difference is that I don't connect to razorback (aMule automatically connects to kalbum.pl :D )
Title: Re: Connection to Kad lost
Post by: Bourdieu on January 18, 2006, 09:37:28 PM: I was connected during the last 2 days, and I didn't experienced any disconnection from Kad network.

Maybe, the problem was due to a temporary congestion on the kad network.
Title: Re: Connection to Kad lost
Post by: lao on January 23, 2006, 07:19:43 PM: Could you please try this patch and see if it fixes this issue for you.

http://www.amule.org/abugs/file_download.php?file_id=118&type=bug
Title: Applied the patch
Post by: acehigh on January 24, 2006, 12:13:34 AM: I applied the patch on the CVS20060117, and relaunched amuled.
Now we'll see.

@Kry: As I already asked before, does it make sense to ask you to put some sort of debug info when it comes to disconnect from kad? Would it be helpful for you if we post logfiles when the disconnect happens?
Title: Re: Connection to Kad lost
Post by: lionel77 on January 24, 2006, 02:14:08 AM: Guys, just a quick question:
If you are using both ed2k and Kad do you get the impression that your download rates drop significantly once Kad is knocked out or do they pretty much stay the same?
Title: RE: Applied the patch
Post by: Kry on January 24, 2006, 08:00:40 AM: Quote
Originally posted by acehigh
@Kry: As I already asked before, does it make sense to ask you to put some sort of debug info when it comes to disconnect from kad? Would it be helpful for you if we post logfiles when the disconnect happens?

I know why it happens. I just don't see why would it happen.
Title: Re: Connection to Kad lost
Post by: Aethereal on January 24, 2006, 10:26:54 AM: Just my 2 cent, and I am not sure is totally relevant, but I had this kind of behaviour:

First of all, kad contacts list gone empty (I do not know why).
It should try to connect using the local ed2k contacts as sources, but after some hours it disconnect and don't try again to reconnect

So I tried, while aMuled was running, to replace nodes.dat with a fresh new one downloaded with a wget while in the .aMule dir, then I used aMuleGUI/Win32 to trigger the "Connect". It then connected and kept connected for days:

Code: [Select]
2006-01-13 11:07:17: Read 0 Kad contacts 2006-01-13 11:07:17: Kad started. 2006-01-13 11:07:27: Kad stopped. 2006-01-13 11:07:27: Connected to Kad (firewalled) 2006-01-13 11:08:50: Connected to Kad (ok) 2006-01-13 13:48:15: Disconnected from Kad --- Manually overwritten nodes.dat 2006-01-16 08:59:01: New external connection accepted 2006-01-16 08:59:01: Connecting client: amule-remote 0x0001 2006-01-16 08:59:01: Access granted. 2006-01-16 08:59:12: Read 185 Kad contacts 2006-01-16 08:59:12: Kad started. --- Kad was on until I reboot on 2006-01-24 09:07:40
When I rebooted, aMuled readed only 93 Kad contacts out of the 185 I had when I substituted the nodes.dat... I suppose the problem is that keeping this behaviour it does soon go out of "live" contacts and stop trying to reconnect (not having nodes to boot from).
Of course there are no "Static" servers on Kad to boot from, but looks like it prune them from the list much more faster than it adds them (it adds them, right?)

Hope this would be of help troubleshoting...
Title: Re: Connection to Kad lost
Post by: Aethereal on January 25, 2006, 08:43:50 AM: Hmm... same exact behaviour after the boot: Kad lasts a couple hours, then give a "Disconnected from Kad" and don't re-connect until I tell it to do via aMuleGUI, then looks like it stay on forever (or at least very long...)

Code: [Select]
2006-01-24 10:16:55: Read 93 Kad contacts 2006-01-24 10:16:56: Kad started. 2006-01-24 10:16:57: Kad stopped. 2006-01-24 10:16:57: Connected to Kad (firewalled) 2006-01-24 10:17:02: Connected to Kad (ok) 2006-01-24 11:17:03: Connected to Kad (firewalled) 2006-01-24 11:17:03: Connected to Kad (ok) 2006-01-24 13:42:01: Disconnected from Kad 2006-01-24 14:22:10: New external connection accepted 2006-01-24 14:22:10: Connecting client: amule-remote 0x0001 2006-01-24 14:22:10: Access granted. 2006-01-24 14:22:20: Read 182 Kad contacts 2006-01-24 14:22:20: Kad started. 2006-01-24 14:25:40: External connection closed. still on at: 2006-01-25 07:26:08
This time the kad contacts count increased without me re-downloading the nodes.dat as before... I'm even more puzzled!

Anyone tried the same?
Title: Re: Connection to Kad lost
Post by: Radek on January 25, 2006, 03:03:27 PM: I tried to reconnect from amulegui/Win32, too.

In the log I could read "Kad started", but never "Connected to Kad (*)".
And in amulegui "Kad: Off" never disappeared.
Title: RE: Applied the patch
Post by: acehigh on January 25, 2006, 07:15:05 PM: Nothing changed by appliyng the patch (as I supposed, and after the last posts).

@Kry:
You know why it happens. You just don't see why would it happen.
? That's why logfiles are for! :)
Title: Re: Connection to Kad lost
Post by: Kry on January 25, 2006, 11:02:41 PM: Logfile would tell me no packets so kad shutdown. nothing more, nothing less.

I might add some massive debug at some time, but right now I'm way more busy doing other things.
Title: RE: Applied the patch
Post by: lao on January 26, 2006, 09:20:52 AM: Quote
Originally posted by acehigh
Nothing changed by appliyng the patch (as I supposed, and after the last posts).

By this you mean that after re-compiling *and* installing with that patch applied, Kad still loses it's connection and does not reconnect? That's unexpected..

Are you sure you have Reconnect enabled? What does grep -rn Reconnect ~/.aMule/*conf show?

Try doing a gmake clean first and see if it still persists.

Put

Code: [Select]
[Debug] Cat_Kademlia\ Main\ Thread=1 into amule.conf and set VerboseDebug=1, that should log any dis- or reconnect, even for the daemon (i think?).
Title: RE: Applied the patch
Post by: Kry on January 26, 2006, 02:22:16 PM: Quote
Originally posted by lao
Quote
Originally posted by acehigh
Nothing changed by appliyng the patch (as I supposed, and after the last posts).

By this you mean that after re-compiling *and* installing with that patch applied, Kad still loses it's connection and does not reconnect? That's unexpected..

No it isnt', as they lose the contacts.
Title: RE: Applied the patch
Post by: lao on January 29, 2006, 12:24:32 PM: Quote
Originally posted by Kry

No it isnt', as they lose the contacts.

Well, I didn't expect it :) Why do they lose all the contacts so fast?

This is what I'm seeing (w/ patch) after roughly 2 days uptime.

Code: [Select]
2006-01-29 12:13:02: Kademlia Main Thread: Stopping Kademlia 2006-01-29 12:13:03: Kademlia Main Thread: Starting Kademlia 2006-01-29 12:13:04: Read 200 Kad contacts
Title: A logfile
Post by: acehigh on January 31, 2006, 07:21:28 PM: Logfile of what happened in my amuled, having enabled the debug as told by lao, and patched the CVS 20060117 with the patch.
When Kad gets disconnected, it reconnects,but i never see the connected to kad (ok) feedback.
Kry is right (obviously :) )

Code: [Select]
2006-01-30 21:10:05: Creditfile loaded, 6100 clients are known 2006-01-30 21:10:05: Loading ipfilter.dat files. 2006-01-30 21:10:05: Loaded 0 IP-ranges from 'ipfilter.dat'. 0 malformed lines were discarded. 2006-01-30 21:10:05: Loaded 0 IP-ranges from 'ipfilter_static.dat'. 0 malformed lines were discarded. 2006-01-30 21:10:05: Loading server.met file: /xxxxxxxxxxxxxxxxx/server.met 2006-01-30 21:10:05: 225 servers in server.met found 2006-01-30 21:10:05: Found 1 part files 2006-01-30 21:10:05: *** TCP socket (ECServer) listening on 0.0.0.0:nnnn 2006-01-30 21:10:05: MuleUDPSocket: Created Server UDP-Socket at port pppp 2006-01-30 21:10:05: MuleUDPSocket: Created Client UDP-Socket at port pppp 2006-01-30 21:10:05: Found 18 known shared files 2006-01-30 21:10:05: Connecting 2006-01-30 21:10:05: Servers: Trying to connect 2006-01-30 21:10:05: Connecting to xxxxxxxxxxxxx (nnn.nnn.nnn.nnn - nnn.nnn.nnn.nnn:pppp) 2006-01-30 21:10:05: Kademlia Main Thread: Starting Kademlia 2006-01-30 21:10:05: AICH Thread: Syncronization thread started. 2006-01-30 21:10:05: AICH Thread: Masterhashes of known files have been loaded. 2006-01-30 21:10:05: AICH Thread: No new files found. 2006-01-30 21:10:05: AICH Thread: Terminated. 2006-01-30 21:10:05: Read 0 Kad contacts 2006-01-30 21:10:06: webserver running on pid 31512 2006-01-30 21:10:06: General: CamuleDaemonApp::OnRun() 2006-01-30 21:10:06: New external connection accepted 2006-01-30 21:10:06: Connected to xxxxxxxxxxxxxx (nnn.nnn.nnn.nnn:pppp) 2006-01-30 21:10:06: Kad started. 2006-01-30 21:10:06: Connecting client: aMuleweb CVS 2006-01-30 21:10:06: Access granted. 2006-01-30 21:10:07: Servers: Connected 2006-01-30 21:10:07: Connection established on: xxxxxxxxxxxxxx 2006-01-30 21:10:07: Connected to xxxxxxxxxxxxx with HighID 2006-01-30 21:10:07: New clientid is xxxxxxxxxxxxx 2006-01-30 21:10:07: ServerMessage: server version .... ..... removed some server message 2006-01-30 21:10:07: Received 4 new servers 2006-01-30 21:10:07: Saving of server-list completed. 2006-01-30 21:10:12: Kad stopped. 2006-01-30 21:10:12: Connected to Kad (firewalled) 2006-01-30 21:10:18: Connected to Kad (ok) 2006-01-30 21:19:11: CFile: Error when opening file (xxxxxxxxxxxxxxxxxx): No such file or directory <--- Ok. Really removed when amuled was down 2006-01-30 21:19:11: Failed to open file (xxxxxxxxxxxxxxxxxxxxxx), removing from list of shared files. 2006-01-30 21:19:46: New external connection accepted 2006-01-30 21:19:46: Connecting client: amule-remote 0x0001 2006-01-30 21:19:46: Access granted. 2006-01-30 21:19:52: Found 8 known shared files 2006-01-30 21:19:52: AICH Thread: Syncronization thread started. 2006-01-30 21:19:52: AICH Thread: Masterhashes of known files have been loaded. 2006-01-30 21:19:52: AICH Thread: No new files found. 2006-01-30 21:19:52: AICH Thread: Terminated. 2006-01-30 21:19:59: External connection closed. 2006-01-30 21:47:11: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-30 22:08:50: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (34) packet in processPublishRequest 2006-01-30 22:10:15: Connected to Kad (firewalled) 2006-01-30 22:10:15: Connected to Kad (ok) 2006-01-30 22:13:23: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-30 23:23:10: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-30 23:28:09: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (34) packet in processPublishRequest 2006-01-30 23:32:46: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-30 23:35:06: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-30 23:39:39: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-30 23:54:10: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-30 23:57:40: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-30 23:58:50: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:13:42: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:18:19: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:21:06: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:28:07: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:29:10: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:34:16: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:38:16: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:44:56: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:46:06: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:47:00: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:51:46: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:55:06: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 00:55:09: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 01:04:15: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 01:07:26: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 01:11:46: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 01:19:46: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 01:22:13: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh <- the same than before 2006-01-31 01:22:15: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 01:23:11: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh <- the same than before 2006-01-31 01:25:30: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (34) packet in processPublishRequest 2006-01-31 01:26:03: Hasher: Creating new thread. 2006-01-31 01:26:03: Hasher: Starting to create MD4 and AICH hash for file: 004.part 2006-01-31 01:26:18: Hasher: Finished hashing file: 004.part 2006-01-31 01:26:18: Hasher: No files on queue, stopping thread. 2006-01-31 01:26:18: Hasher: A thread has died. 2006-01-31 01:26:18: Suspending upload of file: aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 2006-01-31 01:26:18: AICH Thread: Syncronization thread started. 2006-01-31 01:26:18: AICH Thread: Masterhashes of known files have been loaded. 2006-01-31 01:26:18: AICH Thread: No new files found. 2006-01-31 01:26:18: AICH Thread: Terminated. 2006-01-31 01:26:18: Resuming uploads of file: aaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 2006-01-31 01:26:18: Finished downloading: fffffffffffffffffffffffffffffffffffffff 2006-01-31 01:31:56: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 01:34:44: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 01:37:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 01:45:52: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 01:46:52: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 01:47:52: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:01:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:02:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:03:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:04:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:05:34: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:06:34: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:08:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:10:53: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:11:53: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:12:53: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:13:45: Kademlia Main Thread: Unhandled exception on Kad processPacket 2006-01-31 02:14:13: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:15:15: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:16:13: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:17:15: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:18:13: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:19:13: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:19:39: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong type (0x00) in processKademliaRequest 2006-01-31 02:42:13: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:47:13: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 02:56:46: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (34) packet in processPublishRequest 2006-01-31 03:21:14: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:23:14: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:24:38: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:25:55: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:26:54: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:27:46: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 03:28:54: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:29:54: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:30:54: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:32:56: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:34:54: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:38:51: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 03:40:54: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 03:42:54: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 04:21:35: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong size (35) packet in processPublishRequest 2006-01-31 05:27:37: Kademlia Main Thread: Stopping Kademlia 2006-01-31 05:27:37: Disconnected from Kad 2006-01-31 05:27:37: Kademlia Main Thread: Starting Kademlia 2006-01-31 05:27:38: Read 185 Kad contacts 2006-01-31 06:27:37: Kad started. <- NEVER HAD: CONNECTED TO KAD (firewalled or ok) 2006-01-31 13:00:10: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:02:20: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:16:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:17:40: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:18:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:19:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:20:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:22:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:24:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:25:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:26:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:27:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:33:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:37:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:39:31: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:40:31: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:41:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:42:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:45:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:46:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:49:42: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:54:02: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:56:02: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:57:02: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 13:58:12: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 14:00:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 14:02:33: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 14:04:53: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 14:13:33: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 14:52:11: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 14:57:06: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 14:59:06: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:01:06: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:20:46: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:21:46: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:25:06: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:25:56: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:26:08: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:27:06: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:31:12: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:49:56: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:51:56: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:53:57: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 15:57:17: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 16:28:01: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 16:30:03: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 16:45:30: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 16:47:29: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 16:49:30: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 16:52:33: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 16:59:54: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:04:31: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:05:30: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:06:30: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:08:31: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:11:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:14:11: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:17:51: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:19:50: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:21:32: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:23:00: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 17:25:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:22:04: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:23:35: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:35:56: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:37:22: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:40:16: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:41:15: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:42:17: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:43:16: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:44:16: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:46:45: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:47:46: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:48:46: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:50:26: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:50:43: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:53:17: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:54:17: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:58:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 18:59:41: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 19:00:26: Hashset requested for unknown file: hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 2006-01-31 19:00:32: New external connection accepted 2006-01-31 19:00:32: Connecting client: amule-remote 0x0001 2006-01-31 19:00:32: Access granted. 2006-01-31 19:00:40: External connection closed. 2006-01-31 19:01:00: New external connection accepted 2006-01-31 19:01:00: Connecting client: amule-remote 0x0001 2006-01-31 19:01:00: Access granted. 2006-01-31 19:01:06: Found 7 known shared files 2006-01-31 19:01:06: AICH Thread: Syncronization thread started. 2006-01-31 19:01:07: AICH Thread: Masterhashes of known files have been loaded. 2006-01-31 19:01:07: AICH Thread: No new files found. 2006-01-31 19:01:07: AICH Thread: Terminated. 2006-01-31 19:01:10: External connection closed. 2006-01-31 19:03:02: External connection closed. 2006-01-31 19:03:26: New external connection accepted 2006-01-31 19:03:26: Connecting client: aMulecmd CVS 2006-01-31 19:03:26: Access granted. 2006-01-31 19:03:31: ExternalConn: shutdown requested 2006-01-31 19:03:31: Kademlia Main Thread: Stopping Kademlia 2006-01-31 19:03:32: Kad stopped. 2006-01-31 19:03:32: Disconnected from ED2K 2006-01-31 19:03:32: MuleUDPSocket: Shutting down Client UDP-Socket 2006-01-31 19:03:32: MuleUDPSocket: Shutting down Server UDP-Socket
p.s: Sorry for the lenght, but can someone tell me how to reduce the size of the code window?
Title: Re: Connection to Kad lost
Post by: inhumanity on February 02, 2006, 02:56:20 AM: so are you saying kad reconnects but just dosnt display that with the message "Connected to Kad (ok)" ?
Title: Re: Connection to Kad lost
Post by: acehigh on February 04, 2006, 06:46:51 AM: I want to say the same thing than radek:

Quote
Originally posted by Radek
I tried to reconnect from amulegui/Win32, too.

In the log I could read "Kad started", but never "Connected to Kad (*)".
And in amulegui "Kad: Off" never disappeared.

but with the extra "starting kad" caused by the patch. With or without the patch i always see the number of contacts to zero after kad disconnects the first time. Before the first (and the only) disconnection, the number of contacts increases from zero up to about 900 and stays there for some time (hours).
Title: Re: Connection to Kad lost
Post by: Pinky's Brain on February 04, 2006, 10:19:03 PM: It's a bit weird ... AFAICS if KAD is shut down a restart should go exactly the same as when amuled gets started, but somehow it isn't. Shrug, I guess there are some uninitialized variables left which work by accident on first start.
Title: Re: Connection to Kad lost
Post by: Kry on February 04, 2006, 10:21:11 PM: No, there isn't.
Title: Re: Connection to Kad lost
Post by: lao on February 05, 2006, 09:16:39 AM: Quote
Originally posted by acehigh
With or without the patch i always see the number of contacts to zero after kad disconnects the first time. Before the first (and the only) disconnection, the number of contacts increases from zero up to about 900 and stays there for some time (hours).

Well, it does seem as if for some reason the reconnect fails on your end. Unfortunately, after running the daemon w/patch for a few days, I can't reproduce this.

But as your log shows, you still have enough contacts when the reconnect is effected. So I suppose that whatever causes your Kad connection to stagnate for 20 minutes on your end, also prevents the reconnect from working.

With that in mind, try the following patch (applies on top of the previous one).

http://www.amule.org/abugs/file_download.php?file_id=130&type=bug

Note: UNTESTED

It should effectively retry the reconnect every 15 or so minutes, until it succeeds.

HTH,
lao
Title: Re: Connection to Kad lost
Post by: Kry on February 05, 2006, 04:31:29 PM: I keep betting on UDP port dying.
Title: Re: Connection to Kad lost
Post by: lao on February 05, 2006, 05:23:16 PM: Hmm, I've frequently seen amule (which often runs on idle prio here) transfers starving when some other CPU hungry process is running for longer periods of time. Would that fit?
Title: Re: Connection to Kad lost
Post by: Kry on February 05, 2006, 05:36:36 PM: Nope.
Title: Re: Connection to Kad lost
Post by: vdb on February 05, 2006, 05:42:13 PM: Kry, is there any way to check your theory? On a Linux box, that is, is there any way of checking the UDP port death?

Lately I haven't had any disconnects, but then again, aMule crashes invariably in about 2 days, and I am downloading some stuff again.
Title: Re: Connection to Kad lost
Post by: Kry on February 05, 2006, 06:27:32 PM: Activate all debug log for UDP ports, then send a random packet to the UDP port. If it shows in the log as malformed packet, port is ok.
Title: Re: Connection to Kad lost
Post by: EvolutionCrazy on March 03, 2006, 10:58:12 AM: I've the same problem here :(

amuled running only as upload server (never ever download a file)...

i start it, kad stay connected for some time (maybe 2~4hours) and then disconnect....

no way to get it reconnected :(

from amulegui if i click "kad -> disconnected", then "kad -> bootstrap from know clients" in the log i see:
Quote
2006-03-03 10:53:48: Read 199 Kad contacts
2006-03-03 10:53:48: Kad started.

but kad is still died... :(

it's a dedicated server... i've never seen any problem with my connection.... for some time i've also run a CoD2 game server (UDP only) without any problem...
it sounds weird to me a problem on UDP ;(

in amuled i see:
Code: [Select]
WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. Error on CKademliaUDPListener::processPublishRequest received from: 21.58.213.83:4672 No ip match Error on CKademliaUDPListener::processPublishRequest received from: 59.186.48.81:4672 No ip match Invalid Kad tag; type=0xea name=0x00 Error on CKademliaUDPListener::processPublishRequest received from: 189.9.159.82:4672 No ip match
again :(
Quote
2006-03-03 17:11:43: Connected to Kad (firewalled)
2006-03-03 17:11:43: Connected to Kad (ok)
2006-03-03 17:29:54: New external connection accepted
2006-03-03 17:29:54: Connecting client: amule-remote 0x0001
2006-03-03 17:29:54: Access granted.
2006-03-03 17:30:09: External connection closed.
2006-03-03 18:45:53: Disconnected from Kad

Code: [Select]
[xxx@xxx .aMule]$ amuled amuled: OnInit - starting timer Initialising aMule Checking if there is an instance already running... No other instances are running. Loading temp files from /home/xxx/.aMule/Temp. All PartFiles Loaded. ListenSocket: Ok. *** TCP socket (ECServer) listening on 0.0.0.0:4712 *** Server UDP socket (TCP+3) at 0.0.0.0:4665 *** TCP socket (TCP) listening on 0.0.0.0:4662 *** Client UDP socket (extended eMule) at 0.0.0.0:4672 WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. CECSocket error: Input/Output error CECSocket error: Input/Output error
2006/03/05
others errors today:
Code: [Select]
WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. Invalid Kad tag; type=0xc8 name=0x00 Error on CKademliaUDPListener::processPublishRequest received from: 125.175.242.81:50260 No ip match WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. Error on CKademliaUDPListener::processPublishRequest received from: 1.162.233.200:4672 No ip match Error on CKademliaUDPListener::processPublishRequest received from: 106.50.200.83:4672 No ip match WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. Error on CKademliaUDPListener::processPublishRequest received from: 251.183.59.82:4672 No ip match Invalid Kad tag; type=0x82 name=0x00 Error on CKademliaUDPListener::processPublishRequest received from: 89.5.234.82:4672 No ip match WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. CECSocket error: Input/Output error CECSocket error: Input/Output error CECSocket error: Input/Output error CECSocket error: Input/Output error Invalid Kad tag; type=0xdc name=0x00 Error on CKademliaUDPListener::processPublishRequest received from: 244.241.106.212:4672 No ip match Error on CKademliaUDPListener::processPublishRequest received from: 237.1.4.88:4672 No ip match Error on CKademliaUDPListener::processPublishRequest received from: 143.44.46.83:10195 No ip match Error on CKademliaUDPListener::processPublishRequest received from: 209.152.190.84:4672 No ip match Invalid Kad tag; type=0xaa name=0x00 Error on CKademliaUDPListener::processPublishRequest received from: 6.190.6.84:4672 No ip match WARNING! Client UDP-Socket discarded packet due to errors (2) while sending. Error on CKademliaUDPListener::processPublishRequest received from: 66.184.49.83:4672 No ip match Error on CKademliaUDPListener::processPublishRequest received from: 125.95.130.83:4672 No ip match Invalid Kad tag; type=0xfd name=0x00 Error on CKademliaUDPListener::processPublishRequest received from: 16.1.58.83:19607 No ip match WARNING! Client UDP-Socket discarded packet due to errors (2) while sending.
;(
Title: Re: Connection to Kad lost
Post by: TheOnlyFish on April 03, 2006, 09:19:22 AM: It's the same problem here in 2.1.1. Kad is running fine for 48 hours then disconnects. Do I have a problem at my side or is this just the same bug in amuled?
Title: kad
Post by: realcruncher on April 16, 2006, 08:24:49 PM: Somehow the CVS version from 14.4. is still connected to kad after 16hours.
Title: Re: Connection to Kad lost
Post by: Aethereal on April 18, 2006, 09:04:31 AM: Did you had the same problem before, with the same configuration?
I am using CVS from 14 April too but I experienced the same disconnec after some hours as usual.

Looking at the thread, I think I've noticed some point in common between my config and those of other users with the same behaviour, so I'd like to check:

1 - My server has a static private IP address
2 - The server is published on the internet with a static public address
3 - Only the relevant TCP/UDP ports are mapped
4 - The mapping is done by a firewall, not by the local machine (that is unaware of its own public address)
5 - The server mostly "serve" files, but does do downloads only occasionally (there are only 20 files on my DL queue now, all of them rare files of which no parts are available at the moment)

Can someone confirm that?
Title: Re: Connection to Kad lost
Post by: vdb on April 18, 2006, 10:11:55 AM: Yes, I have this same setup. And with 2.1.1 usually the Kad connection stays up for up to 4 days now before giving in. But it still does disconnect after a while:
2006-03-30 20:29:46: Read 146 Kad contacts
2006-03-30 20:29:47: Connected to Kad (firewalled)
2006-03-30 20:29:50: Connected to Kad (ok)
2006-03-30 21:29:47: Connected to Kad (firewalled)
2006-03-30 21:29:48: Connected to Kad (ok)

2006-04-03 12:29:55: Connected to Kad (firewalled)
2006-04-03 12:29:56: Connected to Kad (ok)
2006-04-03 14:12:55: Disconnected from Kad

Perhaps it's because I run the daemon only, never the monolith? And of course, I run the daemon 24*7, until it crashes or it reaches 50% memory usage or it disconnects from Kad, whichever comes first.
Title: Re: Connection to Kad lost
Post by: acehigh on May 02, 2006, 11:53:54 PM: Aethereal requests:

My configuration is similar to yours but:

2) Router has a dynamic IP (no static ip address)

The rest of points are about the same.

I'm stuck to 2.1.1. after I tried many CVSs, but the kad problem still appears.

I'm also involved in ISP blocking some ports (i.e. I got connections with lowID even if the ports were ok, till the day before!!), now solved, but either before or after this problem, kad gets disconnected after a while (2-3 days or so).

P.S. Realcruncher: obviously 16 hours is not enough, just to say.
Title: Re: Connection to Kad lost
Post by: TheOnlyFish on May 03, 2006, 06:35:16 AM: Quote
Originally posted by Kry
I keep betting on UDP port dying.

Activate all debug log for UDP ports, then send a random packet to the UDP port. If it shows in the log as malformed packet, port is ok.

Where do I have to activate the debug log (for the UDP ports) - in the firewall or in amule itself?

(BTW my config is similar to acehigh).
Title: Re: Connection to Kad lost
Post by: phoenix on May 03, 2006, 03:30:52 PM: Quote
Originally posted by TheOnlyFish
Where do I have to activate the debug log (for the UDP ports) - in the firewall or in amule itself?
In aMule:
Preferences->Debugging->Enable Verbose Debugging-Logging plus all UDP related options you want to check.

Cheers!
Title: Re: Connection to Kad lost
Post by: realcruncher on May 15, 2006, 01:11:18 AM: Quote
Originally posted by acehigh

P.S. Realcruncher: obviously 16 hours is not enough, just to say.

But 16 hours is more than 2-3 hours that it lived before. It gets better but still not as good as the monolithic amule.
Title: my 2 cents
Post by: nounounounou on May 16, 2006, 01:16:09 PM: Hi, new on the board but I've been looking quite often here and I decided to register to tell you that:
i'm on mandriva with amuled 2.1.1.
I had the problem fo kad being disconected a few hours after daemon startup...until I change firewall uninstalled shorewall (default firewall under mandriva) and installed instead firestarter...
It's been 1day and 13Hrs now without a kad deconnection... :baby:
has anybody noticed the same as I did?
edit: more details: fw: iptable+firestarter kernel 2.6.16 amule 2.1.1 rpm version for mandriva2006
Title: Re: Connection to Kad lost
Post by: TheOnlyFish on May 16, 2006, 02:18:55 PM: My Firewall is iptables with kernel 2.4.10.

@nounounounou: my amuled sometimes runs for 48 hours before disconnecting. Maybe you have to wait a little longer.

Unfortunately I can't look at the debug messages of amuled. Both amuled and wxwidgets are compiled without the debug option. Recompiling takes some time and I'm in hurry these days.
Title: "Kad: off" after 12h
Post by: Arichy on June 10, 2006, 11:27:59 AM: Running amuled 2.1.2 with router firewall on Debian Sarge.

In the past I had no problems but now after 12h there is "Kad: off" schown.

Seems, Kad is disconnected.

Bootstrapping from known clients does not work with amulegui.

Any possibility to reconnect? Or do I have to restart ?
Title: Re: Connection to Kad lost
Post by: vdb on June 10, 2006, 09:57:39 PM: Sorry, but the only way is to shutdown and restart the amule daemon.

I have noticed that with amuled 2.1.2 on Mandrake Kad usually lives longer than all previous versions, but still it's a bit hit and miss. I have had 6 days of uptime before it disconnected, but also within 14 hours.

I guess it's about UDP on the both the Linux and the router, but it's hard to tell.
Title: Hi all
Post by: acehigh on June 11, 2006, 10:00:40 PM: After some months of usage, and changing OS version, I'm still here with the KAD disconnected problem...
I gave up the CVS to go with some stable versions but 2.1.1 and 2.1.2 have the same problem.

Today I upgraded to 2.1.3. We'll see in a few days.

Ps. Amuled+amulegui is a very very nice thing!!!
Title: Re: Connection to Kad lost
Post by: Arichy on June 23, 2006, 06:45:47 PM: Using amuled 2.1.2 and after some hours or days not only Kad looses connection but also the connection to the server (some of the Donkey) item shows a green and red arrow, while it was all green before. When connecting to another server (also not a donkey one) there is red+green again.

[edit] ok, that means: kad off, ed2k high id. Did not know that one arrow is for one network
Title: Re: Connection to Kad lost
Post by: acehigh on June 23, 2006, 08:19:53 PM: After some days... same problem with 2.1.3
Title: Re: Connection to Kad lost
Post by: Herattic on June 24, 2006, 10:19:08 AM: i also have the same problem with amule 2.1.3
Title: Re: Connection to Kad lost
Post by: vdb on June 24, 2006, 01:21:45 PM: Yep, same here. But.... it seems it takes far longer now to disconnect. The first run with 2.1.3 took over 8 days for Kad to disconnect, now at almost 3 and still going strong.

I also noticed that the memory usage does not increase so much over time anymore, so I think the developers have it almost under control :)

Keep up the good work!
Title: Re: Connection to Kad lost
Post by: Kry on June 24, 2006, 03:09:15 PM: Except there were no changes on Kad for the last versions.
Title: Re: Connection to Kad lost
Post by: vdb on June 25, 2006, 10:46:08 AM: Ah, well, that might explain why right after I wrote that reply Kad disconnected and on the next run stayed connected for only about 8 hours :(

We'll see how it works out.
Title: Re: Connection to Kad lost
Post by: txa2Kj7iw6x on July 21, 2006, 01:04:58 PM: hi, i've the exact same problem described, however, i'm using *emule* under *wine* !
so the bug must be present in the original emule code

hope this helps
Title: Re: Connection to Kad lost
Post by: stefanero on July 21, 2006, 02:59:50 PM: well this bug happens ONLY using amule-daemon, there are no problems with normal aMule
Title: Re: Connection to Kad lost
Post by: vdb on July 25, 2006, 01:48:41 PM: Well, to keep this thread alive, indeed the problem still exists, but the amount of time that the daemon can keep connected to Kad is wildly erratic. In last few weeks, with version 2.1.3 and WxGTK 2.6.3: 8 days, 6 hours, 3 days, 5 days, 12 hours, 2 days, 14 hours, 3 days.

The one thing I hope for is that the "connect Kad" command would actually do something to the daemon in the next release. If this command worked it would be easy to parse the amule.dat file, see if Kad connection is still there and if not send the connect command via amulecmd. All in an easy cron job :)
Title: Re: Connection to Kad lost
Post by: kreegee on July 25, 2006, 08:15:46 PM: yeah, same observations and hopes here :) (with cvs though)
Title: May be a clue
Post by: nickctrl on August 06, 2006, 02:47:16 PM: I'm using amuled + amuleweb since quite a long time now. I'm actually running the 2.1.3 under debian. It's running with a IP filter list of 8 Mb.

I also got the Kad disconnect problem since long time, and every time after no more than 2 hours !

So, 2 days back I tried to disable IpFilter, just to see ...
And wonder of wonders, since then, no Kad disconnection problem :-)

Now in the log I get :
2006-08-04 11:50:42: Connecté à Kad (pare-feu)
2006-08-04 11:50:42: Connecté à Kad (ok)
2006-08-05 01:50:46: Connecté à Kad (pare-feu)
2006-08-05 01:50:47: Connecté à Kad (ok)

I think it's because I'm having a dynamic IP.

But now, I'm connected with ed2k & kad ... and I can do some Kad search without any problem.

Hope it helps :-)
Title: bad news
Post by: nickctrl on August 06, 2006, 04:53:25 PM: The situation was getting better ... but was not solved

After one day I've got the message :

2006-08-06 15:42:36: Déconnecté de Kad

So the problem is not solved by disabling IPFilter ...
It's only occuring much less often ( every fews days instead of every few hours ).
Title: Re: Connection to Kad lost
Post by: EvolutionCrazy on August 07, 2006, 02:01:07 PM: yeah, the problem is still here also in my system...

tired of restarting amuled everyday to get kad connecting i ended setting a cronjob to restart the amuled + amuleweb every 24hours... :(
Title: Re: Connection to Kad lost
Post by: acehigh on October 24, 2006, 08:47:58 PM: The idea of a separate daemon and graphical tool is quite good, but...

I activated a cronjob too.... :-(

sadly, but 24 hours some days is not enough...
I got kad disconnected even after an hour. (stuck to 2.1.3)

For example:
Code: [Select]
2006-10-24 03:16:11: Connected to Kad (ok) 2006-10-24 03:36:05: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:36:38: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:36:43: Kademlia Main Thread: Exception on Kad processPacket: ***NOTE: Received wrong type (0x00) in processKademliaRequest 2006-10-24 03:37:07: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:37:14: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:37:20: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket ... 2006-10-24 03:52:02: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:52:03: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:52:33: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:53:07: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:55:44: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:55:55: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 03:55:56: Kademlia Main Thread: Exception on Kad processPacket: Unknown opcode 11 on CKademliaUDPListener::processPacket 2006-10-24 04:16:25: Kademlia Main Thread: Stopping Kademlia
Title: Re: Connection to Kad lost
Post by: vdb on October 25, 2006, 11:35:15 AM: To get in line with the other posters here I also created a cron job, based on the example in the Wiki. And it does do the job.

I'm looking forward to a new version to see if either Kad is stable or if the 'connect kad' functionality will work. In the last case we won't have to restart the amuled, just issue the command.
Title: Wikii?
Post by: acehigh on October 26, 2006, 10:07:56 PM: Ahhhhh, there is something on the wiki?
I did all by myself..... studied the way to issue cmds with amulecmd -c ,
and after that, I studied cron....
Well... The important is the final result... :)
Title: Re: Connection to Kad lost
Post by: vdb on October 27, 2006, 11:16:23 PM: Hmm, I do experience a problem with my script though:
- I made a check/restart script.
- The script works fine when I run it manually as root.
- I put a call to the script in the crontab of root, it no longer works.

Perhaps somebody can explain this to me, because I have no idea why it would not function if I call it from the crontab? It[ exactly the same command line as when I run it manually, but it just doesn't do a thing. I just don't get it.
Title: Re: Connection to Kad lost
Post by: acehigh on October 27, 2006, 11:52:15 PM: a) Bad idea to run the script as root.
I suggest you create a amule user and launch the script in the crontab of that user.
Much more secure.
b) The path: have you given the correct PATH=/usr/local/bin ... and HOME=/home/user?

My crontab file, for example.. works well:
Code: [Select]
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/bin/X11:/usr/games HOME=/home/userforamule # m h dom mon dow command 00 02 * * * bin/restartamule.sh > bin/restartamule.log 2>&1The bin directory is from the home of 'userforamule'
Hope it helps
Title: Re: Connection to Kad lost
Post by: vdb on October 29, 2006, 11:35:21 AM: If I don't run the script as root I have to start working around the amuled (re)start script, giving out rights to other users. I'd say that running as root makes it more safe in this case.

Anyway, adding the path and home settings appearantly did the job, thanks! It's now working as it should.
Title: Re: Connection to Kad lost
Post by: christophocles on August 27, 2007, 08:10:28 PM: Hi, I noticed it has been several months since the last post on this thread. Has there been any work done to fix this problem?

I am using the shell script to check if KAD is connected and to restart amuled. However, this seems to mess with my router because all network connections passing through the router become very slow after the script restarts amuled a few times. I am thinking amuled is throwing off some weird UDP packets or something. Power-cycling the router seems to make things work properly again.

I haven't seen anyone else complain of this problem. Does this perhaps provide a clue as to whats going on? Obviously, with this router issue, amuled with KAD is no longer an option for me until the bug gets fixed. I really do hope someone with more experience than me can take a look at this problem again.