|
Aiccu failed to start/run after I added a new network card
![[ph]](/s/countries/ph.gif) Shadow Hawkins on Monday, 08 July 2013 06:00:43
Hello,
My Ubuntu server has been originally installed with aiccu using 1 interface card only (eth0). Here, I assigned the IPv6 address provided by sixxs.
Now, I added a new interface card (eth1) and configured it to take the ipv6 address given by sixxs, instead of eth0.
I used radvd and modified it to use eth1 already.
interface eth1 {...}
I couldn't find anything else to update, or maybe I'm missing some files that needed to be updated 'coz Aiccu is failing to start already.
I haven't seen any lines in aiccu.conf that needs to be updated also.
What other files do I need to update?
Any help from you guys is really appreciated.
Thanks!
Aiccu failed to start/run after I added a new network card
Here, I assigned the IPv6 address provided by sixxs.
Which address exactly?
and configured it to take the ipv6 address given by sixxs, instead of eth0.
Which address? One out of the subnet prefix or one out of the tunnel prefix?
I couldn't find anything else to update, or maybe I'm missing some files that needed to be updated 'coz Aiccu is failing to start already.
Check your logs it will show what AICCU does and when it has errors it will report them.
(/var/log/syslog is the best place to look, but error messages might end up in other files depending on syslog configuration)
What other files do I need to update?
As addresses out of the subnet are unrelated to the tunnel addresses, they have no direct relation to AICCU. As such only /etc/network/interface is the file to update and then an ifdown eth1 && ifup eth1 should do the rest.
Aiccu failed to start/run after I added a new network card
Shadow Hawkins on Monday, 08 July 2013 08:05:34
Jeroen Massar wrote:
Which address exactly?
Which address? One out of the subnet prefix or one out of the tunnel prefix?
In my first configuration(where I only have eth0):
auto eth0
iface et0 inet dhcp
auto eth0
iface eth0 inet6 static
address 2401:e800:104::1
netmask 48
endpoint 2401:e800:104:a1::1
ttl 64
pre-up modprobe ipv6
up echo 1 > /proc/sys/net/ipv6/conf/all/forwarding
up echo 0 > /proc/sys/net/ipv6/conf/all/autoconf
After I added eth1:
auto eth0
iface et0 inet dhcp
auto eth1
iface eth1 inet6 static
address 2401:e800:104::1
netmask 48
endpoint 2401:e800:104:a1::1
ttl 64
pre-up modprobe ipv6
up echo 1 > /proc/sys/net/ipv6/conf/all/forwarding
up echo 0 > /proc/sys/net/ipv6/conf/all/autoconf
Check your logs it will show what AICCU does and when it has errors it will report them.
(/var/log/syslog is the best place to look, but error messages might end up in other files depending on syslog configuration)
/var/log/syslog file shows:
Couldn't resolve host tic.sixxs.net, service 3874
Couldn't connect to the TIC server tic.sixxs.net
Couldn't retrieve first tunnel for the above reason, aborting
/var/log/aiccu.log shows:
Starting SixXs Automatic IPv6 Connectivity Client Utility
Checking access to tic.sixxs.net
.
.
.
.
FAIL
Stopping SixXs Automatic IPv6 Connectivity Client Utility
As addresses out of the subnet are unrelated to the tunnel addresses, they have no direct relation to AICCU. As such only /etc/network/interface is the file to update and then an ifdown eth1 && ifup eth1 should do the rest.
I may have configured the interfaces incorrectly but I'm uncertain which part I did wrong.
Aiccu failed to start/run after I added a new network card
auto eth0 iface et0 inet dhcp
Note that there is a 'h' missing from eth0 there, that can cause all kinds of issues, for instance you not having IPv4 and thus not having DNS to resolve the below label from.
This shows that just checkin configurations is not always a good thing, checking the running information is a good idea ("ifconfig" or "ip link show")
ttl 64
Should not be a problem, but normally on Ethernet interfaces one does not have to set a TTL.
pre-up modprobe ipv6
The OS/distro will take care of this, thus it is not needed.
Couldn't resolve host tic.sixxs.net, service 3874
That shows that your DNS server is not able to resolve tic.sixxs.net. Check what is located in /etc/resolv.conf likely caused by not having eth0 up in the first place.
Aiccu failed to start/run after I added a new network card
Shadow Hawkins on Monday, 08 July 2013 09:13:58
Jeroen Massar wrote:
Note that there is a 'h' missing from eth0 there, that can cause all kinds of issues, for instance you not having IPv4 and thus not having DNS to resolve the below label from.
I'm sorry, I mispelled it in my post. But it is actually eth0 in the interfaces file
ttl 64
Should not be a problem, but normally on Ethernet interfaces one does not have to set a TTL.
pre-up modprobe ipv6
The OS/distro will take care of this, thus it is not needed.
> Couldn't resolve host tic.sixxs.net, service 3874
That shows that your DNS server is not able to resolve tic.sixxs.net. Check what is located in /etc/resolv.conf likely caused by not having eth0 up in the first place.
I have done ifconfig and the results show that:
--eth0 ---
inet addr: 10.x.x.x (provided by our domain)
inet6 addr: fe80::219:x* Scope:Link
--eth1--
inet6 addr: fe80::20a:x* Scope:Link
inet6 addr: 2401:e800:104::1/48 Scope:Global
I'm kind of confused.
My IPv4 connection is via eth0 (10.x.x.x)
In /etc/resolv.conf, I'm using our domain's nameserver: 10.x.x.x
So only if nothing's wrong with the nameserver that I am using, tic.sixxs.net would be resolved and aiccu will be running already, right?
Thanks much for the reply Jeroen :) :)
Aiccu failed to start/run after I added a new network card
My IPv4 connection is via eth0 (10.x.x.x)
Is it "working"?
In /etc/resolv.conf, I'm using our domain's nameserver: 10.x.x.x
Is that one working? What are the results of:
dig www.sixxs.net
dig tic.sixxs.net
dig @10.x.x.x www.sixxs.net
dig @10.x.x.x tic.sixxs.net
where 10.x.x.x is your nameserver?
So only if nothing's wrong with the nameserver that I am using, tic.sixxs.net would be resolved and aiccu will be running already, right?
If the DNS resolver in /etc/resolv.conf works, then tic.sixxs.net should be able to be resolved by AICCU yes.
Aiccu failed to start/run after I added a new network card
Shadow Hawkins on Monday, 08 July 2013 11:17:27
Jeroen Massar wrote:
> My IPv4 connection is via eth0 (10.x.x.x)
Is it "working"?
Yes. I tried running ntpdate just to check if ntp is still working (if the port is still open). And I tried pinging other IPv4 addresses in our domain.
> In /etc/resolv.conf, I'm using our domain's nameserver: 10.x.x.x
Is that one working? What are the results of:
Before I did these commands, I tried to do the following:
1. Stopped: radvd, wide-dhcpv6-server, bind9
2. Started aiccu (and it didi start, though i'm not sure why)
3. Started the services again: bind9, wide-dhcp-server, radvd
Then I performed the above commands. Results were:
dig www.sixxs.net
dig tic.sixxs.net
dig @10.x.x.x www.sixxs.net
dig @10.x.x.x tic.sixxs.net
where 10.x.x.x is your nameserver?
dig www.sixxs.net
; <<>> Dig 9.8.1-P1 <<>> www.sixxs.net
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 51279
;; flags: qr rd ra; QUERY: 1, ANSWER: 5, AUTHORITY: 0, ADDITIONAL: 0
;; QUESTION SECTION:
;www.sixxs.net IN A
;; ANSWER SECTION:
www.sixxs.net. 1899 IN CNAME nginx.sixxs.net.
nginx.sixxs.net. 300 IN A 213.197.30.67
nginx.sixxs.net. 300 IN A 213.197.27.252
nginx.sixxs.net. 300 IN A 94.75.219.73
nginx.sixxs.net. 300 IN A 38.229.76.3
;; Query time: 375 msec
;; SERVER: 10.191.23.0#53(10.191.23.0)
;; WHEN: Mon Jul 8 18:08:34 2013
;; MSG SIZE rcvd : 115
dig tic.sixxs.net
; <<>> Dig 9.8.1-P1 <<>> tic.sixxs.net
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 24399
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0
;; QUESTION SECTION:
;tic.sixxs.net IN A
;; ANSWER SECTION:
tic.sixxs.net. 1707 IN A 94.75.219.73
;; Query time: 1 msec
;; SERVER: 10.191.23.0#53(10.191.23.0)
;; WHEN: Mon Jul 8 18:17:34 2013
;; MSG SIZE rcvd : 47
dig @10.191.23.0 www.sixxs.net
; <<>> Dig 9.8.1-P1 <<>> @10.191.23.0 www.sixxs.net
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 31549
;; flags: qr rd ra; QUERY: 1, ANSWER: 5, AUTHORITY: 0, ADDITIONAL: 0
;; QUESTION SECTION:
;www.sixxs.net IN A
;; ANSWER SECTION:
www.sixxs.net. 1139 IN CNAME nginx.sixxs.net.
nginx.sixxs.net. 299 IN A 213.197.30.67
nginx.sixxs.net. 299 IN A 213.197.27.252
nginx.sixxs.net. 299 IN A 94.75.219.73
nginx.sixxs.net. 299 IN A 38.229.76.3
;; Query time: 390 msec
;; SERVER: 10.191.23.0#53(10.191.23.0)
;; WHEN: Mon Jul 8 18:21:14 2013
;; MSG SIZE rcvd : 115
dig @10.191.23.0 tic.sixxs.net
; <<>> Dig 9.8.1-P1 <<>> @10.191.23.0 tic.sixxs.net
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 32871
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0
;; QUESTION SECTION:
;tic.sixxs.net IN A
;; ANSWER SECTION:
tic.sixxs.net. 1121 IN A 94.75.219.73
;; Query time: 0 msec
;; SERVER: 10.191.23.0#53(10.191.23.0)
;; WHEN: Mon Jul 8 18:27:40 2013
;; MSG SIZE rcvd : 47
Weird, aiccu just run after I stopped the services mentioned above and started aiccu ahead of them.
I tried rebooting it again and same thing happened. (aiccu failed again)
And I don't know what particular configuration I need to modify for this. :(
Aiccu failed to start/run after I added a new network card
3. Started the services again: bind9, wide-dhcp-server, radvd
You are configuring eth0 with dhcp, but you also have a DHCP server running?
And the BIND server, is that a recursive/resolver or an authoritive server? And is it the one you are using or another one?
You might also want to check what /etc/resolv.conf states before and after restarting these services; I would not be half-surprised if BIND updates it.
;; SERVER: 10.191.23.0#53(10.191.23.0)
Side-note: Using .0 in an IP like 10.191.23.0 is a bad idea, unless you have configured your network to be a non-/24 (and/or other prefix size where .0 is not the subnet IP; and all equipment is CIDR compliant, noting a lot of it is not) you are very likely asking for a vague random issues. It is very heavily suggested to not use .0.
Aiccu failed to start/run after I added a new network card
Shadow Hawkins on Monday, 08 July 2013 13:05:10You are configuring eth0 with dhcp, but you also have a DHCP server running?
Is this not gonna work or is this not valid?
- eth0 is configured with dhcp (to have an IPv4 address that can connect me to our domain and the internet)
- I am trying to use wide-dhcpv6-server (configured in eth1) so I can configure ipv6 addresses to client PCs (connected via switch) either in DHCP stateful/stateless mode
Is this setup not gonna work?
And the BIND server, is that a recursive/resolver or an authoritive server? Apparently, bind is configured to be an authoritative server based on what I have read:
zone "local.domain" {
type master;
file "local.domain";
}
And is it the one you are using or another one? I only have bind9 (installed this one as part of the DHCPv6 stateful autoconfiguration step that I am following)
Although I tried to uninstall this, upon restart of wide-dhcpv6-client, my client pc could not acquire a global IPv6 address anymore.
You might also want to check what /etc/resolv.conf states before and after restarting these services; I would not be half-surprised if BIND updates it. I will try this later 'coz as of the moment, my server lost its internet connection (adding to my many problems at hand, or maybe this is a side effect o_O).
Side-note: Using .0 in an IP like 10.191.23.0 is a bad idea, unless you have configured your network to be a non-/24 (and/or other prefix size where .0 is not the subnet IP; and all equipment is CIDR compliant, noting a lot of it is not) you are very likely asking for a vague random issues. It is very heavily suggested to not use .0.
- I asked about this from our network admin, our network is configured to be /22 (but I don't know if the equipments are CIDR compliant :| )
you are very likely asking for a vague random issues Oh I'm sorry, please bear with me, I very new to this and I really need help. I'm almost lost with my ipv6 configuration and stuff, but I hope I can figure this out soon...
thank you always.
Aiccu failed to start/run after I added a new network card
Is this not gonna work or is this not valid?
It is just odd and not normally done as all kinds of issues can occur; how do you control your DHCP server when the upstream DHCP server is not there?
- eth0 is configured with dhcp (to have an IPv4 address that can connect me to our domain and the internet)
Thus that is the interesting/relevant part of the setup.
- I am trying to use wide-dhcpv6-server (configured in eth1) so I can configure ipv6 addresses to client PCs (connected via switch) either in DHCP stateful/stateless mode Is this setup not gonna work?
Depends completely on if you configure it properly or not. It can work, but is not typical.
Although I tried to uninstall this, upon restart of wide-dhcpv6-client, my client pc could not acquire a global IPv6 address anymore.
You really have to detail a lot more about your setup, likely there is a very (un)logical configuration error somewhere.
You might also want to check what /etc/resolv.conf states before and after restarting these services; I would not be half-surprised if BIND updates it. I will try this later 'coz as of the moment, my server lost its internet connection (adding to my many problems at hand, or maybe this is a side effect o_O).
Maybe as you are doing DHCP for the server itself and the other DHCP server locked you out because it noticed you are sending out DHCP on it's interface?
The best suggestion I have is to set up your configuration step by step and limit the number of moving parts (dhcp, servers, starting/stopping things while testing) to as much as a minimum as possible.
Also note that when you post here, we'll not see how your system is configured, and as such, we just need to guess; thus provide as many current details as possible, and of course include all the details.
Aiccu failed to start/run after I added a new network card
Shadow Hawkins on Monday, 08 July 2013 08:13:53
Jeroen Massar wrote:
Which address exactly?
Which address? One out of the subnet prefix or one out of the tunnel prefix?
The address I used was from one out of the subnet prefix.
|
![[ch]](/s/countries/ch.gif)
on Monday, 08 July 2013 06:06:37
Posting is only allowed when you are