Hi.

This morning, I tried to upgrade my 2 OpenVPN servers from 2.2.2 to 
2.3.6.  With the clients down, I started the 2.3.6 OpenVPN server. The 
server started fine. I then tried to start the clients, which were 
already running 2.3.6. The clients would hang and seemed unable to 
connect to the server.  On the server, I kept seeing:

Wed Jun  3 10:22:37 2015 us=509679 brayden/130.63.97.125:44636 RESOLVE: 
Cannot resolve host address: netmask: Name or service not known
Wed Jun  3 10:22:37 2015 us=509778 brayden/130.63.97.125:44636 MULTI: 
Learn: 172.16.37.125 -> brayden/130.63.97.125:44636
Wed Jun  3 10:22:37 2015 us=509795 brayden/130.63.97.125:44636 MULTI: 
primary virtual IP for brayden/130.63.97.125:44636: 172.16.37.125

Running 2.2.2 on the server, the "RESOLVE:" error line did not exist.

I tried to enable extra debugging on the server, but really, I couldn't 
see any errors other than the resolve error.

Since I had a very short window where I could do the upgrade, I resorted 
the OpenVPN servers back to 2.2.2.  The clients were still running 
2.3.6.  Everything was fine.

In retrospect, I wish I had enabled logging on the client, and capture 
verbose logs.  Maybe there were additional errors there.

I setup a test 2.3.6 server with the identical config, and a 2.3.6 
client, and the 2.3.6 client connects to the test 2.3.6 server perfectly 
fine, even though the "RESOLVE" error is still generated in the log 
file.  I wish I could understand why it failed this morning, but it 
works now.  That being said, I'd like to understand why the RESOLVE 
error is generated, and how to get rid of it before I try the upgrade to 
2.3.6 again.  The error seems to be complaining about both the host 
address, and the netmask, but it seems to resolve fine in the next line.

Here's a few more details..

server config:

dev tun
proto udp
port 1194
tmp-dir /tmp
writepid /var/run/openvpn-server.pid
crl-verify /xconf/openvpn/pki/crl.pem
mode server
tls-server
ifconfig 172.16.32.26 172.16.32.1
route 172.16.32.0 255.255.240.0
push "route 172.16.0.0 255.255.240.0"
client-connect /xsys/lib/openvpn-tools/connect
keepalive 10 20
user nobody
group nobody
persist-key
persist-tun
status /var/log/openvpn-status
log-append /var/log/openvpn-server
verb 4
mute 20
cipher none
comp-lzo
txqueuelen 1000
daemon

client config:

client
dev tun
proto udp
remote copper.cs.yorku.ca
remote nickel.cs.yorku.ca
remote-random
writepid /var/run/openvpn.pid
nobind
user nobody
group nobody
persist-key
persist-tun
persist-remote-ip
ns-cert-type server
verb 3
mute 20
cipher none
comp-lzo
daemon

connect script:

In short, it returns for a given host "peach":

ifconfig-push peach-vpn2 copper-vpn2 netmask 255.255.240.0

(It's the same script I've been using since OpenVPN 2.0.9 days)

server startup:

openvpn --ca /xconf/openvpn/pki/ca.crt --dh 
/xconf/openvpn/pki/dh2048.pem --cert 
/xconf/openvpn/pki/copper.server.crt --key 
/etc/openvpn/copper.server.key --config 
/xconf/openvpn/config/server.conf.coppy --script-security 3 system

client startup:

openvpn --ca /xconf/openvpn/pki/ca.crt --cert 
/xconf/openvpn/pki/jun48.crt --key /etc/openvpn/jun48.key --config 
/xconf/openvpn/config/client.conf

In addition, on both the servers, /etc/resolv.conf is identical, 
pointing to our DNS.  In addition, /etc/hosts contains all the hostnames 
as well.

Thanks for any help you can provide..

Jason.


------------------------------------------------------------------------------
_______________________________________________
Openvpn-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/openvpn-users

Reply via email to