gnunet-developers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [GNUnet-developers] newbie Q: connecting ..


From: Christian Grothoff
Subject: Re: [GNUnet-developers] newbie Q: connecting ..
Date: Sun, 15 Jun 2003 19:49:48 -0500
User-agent: KMail/1.4.3

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sunday 15 June 2003 07:00 pm, you wrote:
> Hi Christian
>
> Here's more info, apologies if it's too much ..
>
> [Hey, it works! See end .. ]
>
> > Well, what I can see on the Gentoo webpage at
> > http://www.gentoo.org/dyn/pkgs/net-p2p/gnunet.xml
> >
> > is that they have an ancient version of GNUnet,
>
> Weird, that URL lists 0.4.6, which is what I had when I wrote the
> original post, but they're distributing GNUnet v0.5.2a now, so
> the URL is out of date. And I would have expected it to be
> autogenerated when they updated the distributed version. I'll check
> it out and file a bug report if necessary.

Not to mention that 0.5.2a is not quite the latest version :-)

> > Could you also give some more details on the architecture that you're
> > using? Also, the output of "ldd `which gnuned`" will be useful since it
> > can tell us which crypto-library is being used.
>
> I'm on a 650Mhz PIII IBM Thinkpad 600x with 256MB RAM.
>
> address@hidden src $ ldd `which gnunetd`
>          libgnunetutil.so.0 => /usr/lib/libgnunetutil.so.0 (0x41017000)
>          libdl.so.2 => /lib/libdl.so.2 (0x41029000)
>          libm.so.6 => /lib/libm.so.6 (0x4102c000)
>          libnsl.so.1 => /lib/libnsl.so.1 (0x4104e000)
>          libpthread.so.0 => /lib/libpthread.so.0 (0x41063000)
>          libcrypto.so.0.9.6 => /usr/lib/libcrypto.so.0.9.6 (0x410b4000)
>          libc.so.6 => /lib/libc.so.6 (0x41178000)
>          /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40ff7000)
>
> Current symptoms is that this does nothing:
>
> address@hidden jean $ gnunetd
>
> and ps shows no gnunetd.

In ~/.gnunet/logs, you would have been able to find the error-message.

> If I tell it not to detach:
>
> address@hidden jean $ gnunetd -d
> Jun 16 01:05:53 Could not open library libgnunetafs_protocol.so
> (libgnunetafs_protocol.so: cannot open shared object file: No such file or
> directory)!
> Jun 16 01:05:53 __BREAK__
>
> I was a bit surprised that this:
>
> klippie jean # grep -r libgnunetafs_protocol.so /var/log/
>
> didn't show any trace of the first invocation failing.

GNUnet logs to ~/.gnunet/logs, not /var/log, unless gnunet.conf is changed. 
The reason is, that gnunetd may not have write permissions for /var/log.

> > All GNUnet dependencies are listed at
>
> Thanks. In a previous post I'd also found the dependencies to be due to
> doc generation.
>
> I've done
>   ./configure --prefix=/usr/bin
>   ./make
> now. This built libgnunetafs_protocol.so, which I copied into place:
>
> cp
> /home/jean/src/GNUnet-0.5.2a/src/applications/afs/protocol/.libs/libgnuneta
>fs_protocol.so /usr/lib/
>
> and now gnunetd starts up OK. (To avoid copying files around that aren't
> under Gentoo package management I'll still have to figure out why the so
> didn't get built in the first place ..)

Right, that is very likely a packaging error.

> It still doesn't talk though:
>
> address@hidden GNUnet-0.5.2a $ gnunetd -d
> Jun 16 01:20:11 WARNING: received incomplete/invalid helo from http
> Jun 16 01:20:19 WARNING: announcing ourselves pointless: no hosts known
> Jun 16 01:20:19 WARNING: announcing ourselves pointless: no hosts known

Strange, but it may be that you hit a point in time where the hostlisturl was 
just updated on ovmj.org (unlikely, but can theoretically happen). The other 
warnings may be ok -- if they go away after that.

> address@hidden jean $ gnunet-transport-check
> Testing transport(s) udp tcp
> Jun 16 01:23:30 ERROR (Address already in use) binding the TCP listener to
> port 2086. No transport service started.
> ERROR: could not start transport server
> Transport OK, 0ms for 0 messages of size 11 bytes.

You can not run transport-check while gnunetd is running. Check "ps ax" and 
use netstat to see if the port is in use.

> The config file is unchanged from the distro except for s/eth0/ppp0/
>
> Hmm, now it's suddenly more chatty:
>
> address@hidden GNUnet-0.5.2a $ gnunetd -d
> Jun 16 01:20:11 WARNING: received incomplete/invalid helo from http
> Jun 16 01:20:19 WARNING: announcing ourselves pointless: no hosts known
> Jun 16 01:20:19 WARNING: announcing ourselves pointless: no hosts known
> Jun 16 01:26:12 WARNING: makeSesionKeySigned: cannot encrypt sessionkey,
> other host not known!
> address@hidden GNUnet-0.5.2a $ gnunetd -d
> Jun 16 01:26:41 FAILURE: decrypting message from host
> 1A10C4058997A9748D3490F2262C6BB7E2534E4F failed, no sessionkey!
> Jun 16 01:26:41 WARNING: Can not decrypt packet from host
> 1A10C4058997A9748D3490F2262C6BB7E2534E4F: no session key? (1444).
> Jun 16 01:26:44 FAILURE: decrypting message from host
> 1BA51B8DFC0D3B055D3ACB237CA162BD770E46C9 failed, no sessionkey!
> Jun 16 01:26:44 WARNING: Can not decrypt packet from host
> 1BA51B8DFC0D3B055D3ACB237CA162BD770E46C9: no session key? (1444).
> Jun 16 01:26:49 FAILURE: decrypting message from host
> 1A10C4058997A9748D3490F2262C6BB7E2534E4F failed, no sessionkey!
> Jun 16 01:26:49 WARNING: Can not decrypt packet from host
> 1A10C4058997A9748D3490F2262C6BB7E2534E4F: no session key? (1444).
> Jun 16 01:26:49 FAILURE: decrypting message from host
> 1A10C4058997A9748D3490F2262C6BB7E2534E4F failed, no sessionkey!
> Jun 16 01:26:49 WARNING: Can not decrypt packet from host
> 1A10C4058997A9748D3490F2262C6BB7E2534E4F: no session key? (1444).
> Jun 16 01:27:30 WARNING: received incomplete/invalid helo from http

I wonder if some of your errors may come from old files lying around in 
~/.gnunet/ from the previous 0.4.x installation. Did you update the 
gnunet.conf file?

> This breaks:
>
> address@hidden jean $ gnunet-tbench
>
> Did not receive the message from gnunetd. Is gnunetd running?
>
> It results in:
>
> address@hidden GNUnet-0.5.2a $ gnunetd -d
> Jun 16 01:28:36 assertion failed: strlen(hex) is not 40
> Jun 16 01:28:36 __BREAK__

Is this 0.5.2a or 0.5.4? I don't really remember which bugs were fixed in 
between :-). Anyway, this is definitely a GNUnet bug. 

> Heh, more new messages, it looks like *something*'s coming through?
>
> address@hidden GNUnet-0.5.2a $ gnunetd -d
> Jun 16 01:29:26 FAILURE: decrypting message from host
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1 failed, no sessionkey!
> Jun 16 01:29:26 WARNING: Can not decrypt packet from host
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1: no session key? (1444).
> Jun 16 01:30:14 WARNING: could not download hostlist, host www.ovmj.org
> unknown Jun 16 01:30:25 WARNING: CRC failed from
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1, msg ignored (wrong skey?)
> Jun 16 01:30:28 WARNING: CRC failed from
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1, msg ignored (wrong skey?)
> Jun 16 01:30:29 WARNING: CRC failed from
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1, msg ignored (wrong skey?)
> Jun 16 01:30:45 WARNING: CRC failed from
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1, msg ignored (wrong skey?)
> Jun 16 01:30:48 WARNING: CRC failed from
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1, msg ignored (wrong skey?)
> Jun 16 01:31:20 WARNING: CRC failed from
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1, msg ignored (wrong skey?)
> Jun 16 01:31:21 WARNING: CRC failed from
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1, msg ignored (wrong skey?)
> Jun 16 01:31:24 WARNING: CRC failed from
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1, msg ignored (wrong skey?)

You will get CRC errors if certain GNUnet messages were lost on your 
connection. Since you seem to be droping 50%, this could actually be 
perfectly normal.

> address@hidden jean $ gnunet-peer-info
> 7E808C65630219C56AD9FE25FF97AF6C8D05A7CD at address 62.131.97.197:2086
> (UDP) CB6DB904F1C519D30317470193B701D014703004 at address 128.10.19.51:2086
> (UDP) 1BA51B8DFC0D3B055D3ACB237CA162BD770E46C9 at address
> 12.222.130.93:2086 (TCP) 2F7A9C9A4D62E128EDD5D444346F129F9916CAC8 at
> address 80.145.86.4:2086 (UDP) 1A10C4058997A9748D3490F2262C6BB7E2534E4F at
> address 128.211.1.61:2086 (TCP) 7E808C65630219C56AD9FE25FF97AF6C8D05A7CD at
> address 62.131.97.197:2086 (TCP) 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1
> at address 217.10.9.100:2086 (UDP) 4228876A46F655C551E2A3DF3C2035AC26229BA2
> at address 128.46.101.30:2086 (TCP)
> 1BA51B8DFC0D3B055D3ACB237CA162BD770E46C9 at address 12.222.130.93:2086
> (UDP) 9CB8D759542525AE7E06ADD8239150325B81DEB9 at address
> 195.197.251.181:2086 (TCP) 2F7A9C9A4D62E128EDD5D444346F129F9916CAC8 at
> address 80.145.86.4:2086 (TCP) 1A10C4058997A9748D3490F2262C6BB7E2534E4F at
> address 128.211.1.61:2086 (UDP) CB6DB904F1C519D30317470193B701D014703004 at
> address 128.10.19.51:2086 (TCP) 42F91CAB229B3119F80A0D447A12FE867B60895F at
> address 212.90.68.242:2086 (UDP) 42F91CAB229B3119F80A0D447A12FE867B60895F
> at address 212.90.68.242:2086 (TCP)
> 7DC4C7B9C295ADD8D15FC5253457F53DAC20D0E1 at address 217.10.9.100:2086 (TCP)
> 4228876A46F655C551E2A3DF3C2035AC26229BA2 at address 128.46.101.30:2086
> (UDP)
>
> I might mention that I'm on a dialup connection that sucks very hard, but
> very ineffectively:
>
> address@hidden jean $ ping www.google.com
> PING www.google.com (216.239.39.99): 56 octets data
> 64 octets from 216.239.39.99: icmp_seq=0 ttl=48 time=7926.7 ms
> wrong data byte #0 should be 0x1d but was 0x1616 2 ed 3e 36 d5 9 0
>          8 9 a b c d e f 10 11 12 13 14 15 16 17 18 19 1a 1b 1c 1d 1e 1f 20
> 21 22 23 24 25 26 27
>          28 29 2a 2b 2c 2d 2e 2f
> 64 octets from 216.239.39.99: icmp_seq=3 ttl=48 time=7847.3 ms
> wrong data byte #0 should be 0x20 but was 0x1919 2 ed 3e e5 dd 9 0
>          8 9 a b c d e f 10 11 12 13 14 15 16 17 18 19 1a 1b 1c 1d 1e 1f 20
> 21 22 23 24 25 26 27
>          28 29 2a 2b 2c 2d 2e 2f
> [...]
> 302 packets transmitted, 147 packets received, 51% packet loss
>
> Perhaps this is too noisy for gnunet. Let's try a search ..

It should just be slow, but it should always work -- where of course work at a 
99% loss-rate may mean that it takes days to get a search result :-)

> address@hidden jean $ gnunet-search GPL
> gnunet-download -o "COPYING" -- 77DBE6F971D66F641B2262DDCE78CA8FB6815E60
> E34BFD7843D4C8469EFC3DAD260F1F71783E2A87 466DC92 17992
> => GNU General Public License v2 <= (mimetype: unknown)
>
> Hey, cool, it works!
>
> address@hidden jean $ gnunet-download -o "COPYING" --
> 77DBE6F971D66F641B2262DDCE78CA8FB6815E60
> E34BFD7843D4C8469EFC3DAD260F1F71783E2A87 466DC92 17992
> Download at        0 out of    17992 bytes
> Download at     1024 out of    17992 bytes
> Download at     2048 out of    17992 bytes
> Download at     3072 out of    17992 bytes
> Download at     4096 out of    17992 bytes
> Download at     4680 out of    17992 bytes
> Download at     5704 out of    17992 bytes
> Download at     6728 out of    17992 bytes
> Download at     7752 out of    17992 bytes
>
> Thanks for the help :]  I was thrown by gnunetd's debug messages: I thought
> they were all signs of misconfiguration. I didn't realise that they were
> diagnostics and that gnunetd just needed a while to orient itself, as it
> were.

It definitely needs a while to orient itself, it also needs a bit to clean up 
garbage from previous installations. And yes, the diagnostics are very 
verbose :-). 

Glad to hear that it's working.

Christian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE+7RQs9tNtMeXQLkIRAgNxAJkBvBCbF6w/agXmJanvXaY3niumgACeOONj
a041vJeABydtHPhaUXMmCn0=
=q3YF
-----END PGP SIGNATURE-----





reply via email to

[Prev in Thread] Current Thread [Next in Thread]