Hello everyone,

Hope I am writing on the right forum, if not - my apologies.
I am new to Maemo and have no experience with previous devices, so I can't compare behaviour.

I started to play with new and shiny N900 with noble goal to develop Qt based application. One of steps on the way is to establish ssh to N900. I can't say it does not work completely. It works. Somewhat. I installed the latest openssh, 5.1p1-6 maemo5, and that's what I got.

1. I can not connect at all from Hardy Ubuntu laptop via WiFi. Probably Hardy's problem, read about similar problems on forums.

2. I managed to connect to the same Hardy laptop from N900, not always but sometimes

3. SSH within N900 works fine, from user account to root, the server is running.

4. I sometimes can connect via wires from another laptop running Mepis.

[I]ssh -l root 192.168.0.103
The authenticity of host '192.168.0.103 (192.168.0.103)' can't be established.
RSA key fingerprint is ba:6e:e8:d7:a0:c3:3f:57:75:92:4b:b2:ea:37:6b:85.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added '192.168.0.103' (RSA) to the list of known hosts.
root@192.168.0.103's password:


BusyBox v1.10.2 (Debian 3:1.10.2.legal-1osso26+0m5) built-in shell (ash)
Enter 'help' for a list of built-in commands.

Nokia-N900-42-7:~#
Nokia-N900-42-7:~#
Nokia-N900-42-7:~# ls
Nokia-N900-42-7:~# lsls
-sh: lsls: not found

Nokia-N900-42-7:~#
Nokia-N900-42-7:~# lsls
-sh: lsls: not found
Nokia-N900-42-7:~# ls
Nokia-N900-42-7:~#
Nokia-N900-42-7:~# Read from remote host 192.168.0.103: Connection timed out
Connection to 192.168.0.103 closed.

N900 responds very sloooowwwly. This is session of about 12 minutes. Other way around i.e. N900 ->laptop was much better, not perfect either but at least respond was within seconds, not minutes.

5. I was pinging the N900 to keep wifi alive, here is the result.


ping 192.168.0.103 -i 5
PING 192.168.0.103 (192.168.0.103): 56 data bytes
64 bytes from 192.168.0.103: icmp_seq=4 ttl=64 time=81.5 ms
64 bytes from 192.168.0.103: icmp_seq=15 ttl=64 time=2049.5 ms
64 bytes from 192.168.0.103: icmp_seq=27 ttl=64 time=2105.5 ms
64 bytes from 192.168.0.103: icmp_seq=51 ttl=64 time=2122.3 ms
64 bytes from 192.168.0.103: icmp_seq=68 ttl=64 time=13.0 ms
64 bytes from 192.168.0.103: icmp_seq=81 ttl=64 time=2042.3 ms
64 bytes from 192.168.0.103: icmp_seq=91 ttl=64 time=2052.8 ms
64 bytes from 192.168.0.103: icmp_seq=111 ttl=64 time=2119.5 ms
64 bytes from 192.168.0.103: icmp_seq=135 ttl=64 time=2036.8 ms
64 bytes from 192.168.0.103: icmp_seq=147 ttl=64 time=2097.8 ms
64 bytes from 192.168.0.103: icmp_seq=159 ttl=64 time=2057.4 ms
64 bytes from 192.168.0.103: icmp_seq=171 ttl=64 time=2116.3 ms
64 bytes from 192.168.0.103: icmp_seq=172 ttl=64 time=81.9 ms
64 bytes from 192.168.0.103: icmp_seq=183 ttl=64 time=2079.0 ms
64 bytes from 192.168.0.103: icmp_seq=207 ttl=64 time=1073.5 ms
64 bytes from 192.168.0.103: icmp_seq=219 ttl=64 time=1129.9 ms
64 bytes from 192.168.0.103: icmp_seq=231 ttl=64 time=1070.1 ms
64 bytes from 192.168.0.103: icmp_seq=255 ttl=64 time=1086.6 ms
64 bytes from 192.168.0.103: icmp_seq=267 ttl=64 time=1047.4 ms
64 bytes from 192.168.0.103: icmp_seq=279 ttl=64 time=1108.5 ms
64 bytes from 192.168.0.103: icmp_seq=291 ttl=64 time=1064.3 ms
64 bytes from 192.168.0.103: icmp_seq=303 ttl=64 time=1116.7 ms
64 bytes from 192.168.0.103: icmp_seq=339 ttl=64 time=1123.2 ms
64 bytes from 192.168.0.103: icmp_seq=351 ttl=64 time=1081.7 ms
64 bytes from 192.168.0.103: icmp_seq=363 ttl=64 time=1123.5 ms
64 bytes from 192.168.0.103: icmp_seq=375 ttl=64 time=1078.3 ms
64 bytes from 192.168.0.103: icmp_seq=387 ttl=64 time=1033.0 ms
64 bytes from 192.168.0.103: icmp_seq=411 ttl=64 time=1042.6 ms
64 bytes from 192.168.0.103: icmp_seq=423 ttl=64 time=42.0 ms
64 bytes from 192.168.0.103: icmp_seq=435 ttl=64 time=101.2 ms
64 bytes from 192.168.0.103: icmp_seq=447 ttl=64 time=61.5 ms
64 bytes from 192.168.0.103: icmp_seq=459 ttl=64 time=117.8 ms
64 bytes from 192.168.0.103: icmp_seq=471 ttl=64 time=77.0 ms
64 bytes from 192.168.0.103: icmp_seq=483 ttl=64 time=35.9 ms
64 bytes from 192.168.0.103: icmp_seq=495 ttl=64 time=95.3 ms
64 bytes from 192.168.0.103: icmp_seq=507 ttl=64 time=55.3 ms
64 bytes from 192.168.0.103: icmp_seq=513 ttl=64 time=54.1 ms
64 bytes from 192.168.0.103: icmp_seq=525 ttl=64 time=2042.5 ms
64 bytes from 192.168.0.103: icmp_seq=526 ttl=64 time=108.8 ms
64 bytes from 192.168.0.103: icmp_seq=543 ttl=64 time=45.4 ms
64 bytes from 192.168.0.103: icmp_seq=555 ttl=64 time=101.9 ms
64 bytes from 192.168.0.103: icmp_seq=567 ttl=64 time=47.9 ms
64 bytes from 192.168.0.103: icmp_seq=579 ttl=64 time=121.6 ms
64 bytes from 192.168.0.103: icmp_seq=615 ttl=64 time=84.5 ms

--- 192.168.0.103 ping statistics ---
633 packets transmitted, 45 packets received, 92% packet loss
round-trip min/avg/max = 13.0/921.4/2122.3 ms


What is your experience with ssh'ing to N900?

Any suggestion will be much appreciated.


Boris.

Do you have the same packet loss wired? With this kind of packet loss ofcourse you're getting bad network-related performance... probably not only SSH which works shitty.

I am not sure if this is the right way to be measuring network performance. Firstly, how many hops are there to your device - is there a router in the way? Maybe using some tool like iperf might help though that is a bit complicated to install. tcpdump might help too or going the simplest route with ssh -vv.

In any case, the data here is a little too small to serve as a representative sample.

WiFi power-saving causes problems with SSH.
Making connections from computer to tablet often fails because the tablet doesn't answer. Doing some network action on the tablet activates WiFi again and you'll get a response.

For SSH and SCP I open a X-Terminal on the N900, gain root, and start pinging my router. Then I can easily access the tablet from my computer without having to turn off power-saving. When I'm done, I stop the pinging.

The N900 must send the pings to keep the WiFi alive, not the other way round.

I'm not entirely convinced by that: to keep alive the WiFi, having traffic on the interface should be enough. By pinging from an external source the device, you will generate traffic (both inbound and outbound), thus keeping alive the WiFi...but maybe it's different on the N900?

The connection between Mepis laptop and router is wired. The router is within meter and a half from both devices
What bothers me is pretty constant time between responses on Ping, around 2 sec initially, and around 1 sec after. The spikes happened when I activated N900.

That's what I thought while pinging N900 from outside. It did not work. Pinging other way around keeps wifi active, with no packet loss. The time is between 10-40 msec, as opposite to 2 sec. This is probably specific to N900.

Thanks pycage.
It works somewhat better now. Still it is very slow response to commands entered, seconds, sometimes tens of seconds. Do you see the same ?

If you run Mepis laptop on LAN and other laptop on WLAN performance is acceptable? At this point I'd start to do some packet capturing...