Connect ok but I get a lot of econnrefused and ehostunreach

[cmatte]

Hi,
my server works ok since days, but looking at the logs I have lots of repeated lines with these errors:

Sat Feb 26 18:02:49 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:02:53 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:02:57 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:02 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:06 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:10 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:15 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:19 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:23 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:27 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:33 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:37 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:41 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:45 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:48 2011 read UDPv4 [ECONNREFUSED|ECONNREFUSED|ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:48 2011 ip:port write UDPv4 [ECONNREFUSED|ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:48 2011 read UDPv4 [ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:50 2011 ip:port write UDPv4 [ECONNREFUSED|ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:50 2011 read UDPv4 [ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:53 2011 ip:port write UDPv4 [ECONNREFUSED|ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:53 2011 read UDPv4 [ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:55 2011 ip:port write UDPv4 [ECONNREFUSED|ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:55 2011 read UDPv4 [ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:57 2011 ip:port write UDPv4 [ECONNREFUSED|ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:57 2011 read UDPv4 [ECONNREFUSED]: Connection refused (code=146)

This is a Windows pc connecting to the server. It happens on connections or on disconnections. I thought it could be the firewall but also another device with no firewall that is 100% reachable as far as I know creates these logs.
The connection is anyway initiated within seconds from these so the configs are perfectly aligned, I only would like not to see all these logs there because they grow fast and I don't find a sense in these. I'm using no verb option so there isn't a chance I can get the log less verbose and avoid these lines, is there? Do you think there's a problem somewhere? Advices to obtain the wanted result (no such lines even if the "no-problem" persists)?

As always, thanks.

[maikcat]

hi there,


very stupid question...

are you trying to connect from a pc which is in the same network as your server?


please post configs (server/client) and server client OS info.

cheers,

michael.

[janjust]

normally when I see lots of

Sat Feb 26 18:03:45 2011 read UDPv4 [EHOSTUNREACH|EHOSTUNREACH|EHOSTUNREACH]: No route to host (code=148)
Sat Feb 26 18:03:48 2011 read UDPv4 [ECONNREFUSED|ECONNREFUSED|ECONNREFUSED]: Connection refused (code=146)
Sat Feb 26 18:03:48 2011 ip:port write UDPv4 [ECONNREFUSED|ECONNREFUSED]: Connection refused (code=146)

messages it is caused by a bad network connection - the packets to/from the OpenVPN server are dropped somewhere along the way.

You can try switching to 'tcp' mode to see if that helps - in that case you've got a bad router/switch in between which is not passing through UDP packets correctly.

[cmatte]

Yes, I am in the same internal home network of the server when doing these tests. I guess this is an actual problem due to messed up routes, right?
I also just found out that the log of the client is full of these:

Tue Mar 19 07:11:57 1985 TLS Error: Unroutable control packet received from ip:port (si=3 op=P_ACK_V1)
Tue Mar 19 07:11:58 1985 TLS Error: Unroutable control packet received from ip:port (si=3 op=P_CONTROL_V1)

so perhaps the "problem" would be solved connecting from out of the home network? Or using the internal address instead of the external one?

[maikcat]

hi cmatte,

if you want to test your setup in the same internal network dont push routes to your clients,

the ideal is to check your setup from another network (internet..).

cheers,

michael.

[cmatte]

I am having issues.
Connection is established and works ok but after a random time (10h~72h...never more) it falls with the same messages.
[ehostunreach]
Do you think there's a loss of wan connection here that causes these problems?
What else could it be?
Thanks!

[maikcat]

hi there,

>The client is under a MAN and the ip on the internet is not its own, I think for some causes it could be not pingable sometimes even if the connection is up and running. Could this be a real good reason for this?

hmm...

maybe something breaks routing..?

if you do a traceroute while those disconnections
is it shows that routing is correct?

cant think something else right now..

Michael.

[cmatte]

Everything has been explained.
Disconnections are due to real network connectivity and nothing else, I had the proof. Thanks!