Given that I can't install any extra software, any suggestions on how
to fix this ?
Am 22.01.2023 um 10:59:28 Uhr schrieb Adrian:
Given that I can't install any extra software, any suggestions on how
to fix this ?
Make sure the NIC of the VM is connected as a network bridge in VBox.
Show the output of
ip a
ip -6 route show
ip -4 route show
I have two VMs running (one W10/32 one CentOS), both appear to have
the same settings. The W10/32 works fine, I'm posting this from that
VM.
2: enp0s3: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc^^^^
pfifo_fast state DOWN group default qlen 1000
Am 22.01.2023 um 13:25:29 Uhr schrieb Adrian:
I have two VMs running (one W10/32 one CentOS), both appear to have
the same settings. The W10/32 works fine, I'm posting this from that
VM.
Please check the settings.
2: enp0s3: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc^^^^
pfifo_fast state DOWN group default qlen 1000
That means the NIC isn't active. Check VBox if it is "connected". VBox
can virtually disconnect a NIC of VM, like pulling the cable on a
physical machine.
Next question is:
How is your network managed in the virtual OS?
Am 22.01.2023 um 13:25:29 Uhr schrieb Adrian:
I have two VMs running (one W10/32 one CentOS), both appear to have
the same settings. The W10/32 works fine, I'm posting this from that
VM.
Please check the settings.
Next question is:
How is your network managed in the virtual OS?
Now what I think has happened is that Oracle's VBox simply needs to
be told the new router's IP address?
Is the new router IP different to the old router IP?
Now what I think has happened is that Oracle's VBox simply needs to be
told the new router's IP address?
Alternatively, has the physical IP of the machine hosting the VM's
changed? The VMs may still think the host machine still has the old IP
(that was given out by the old router) and simply need to be told what
the host machine's new IP is that is given out by new router?
In message <tqjv4c$70f$1@gioia.aioe.org>, SH <i.love.spam@spam.com> writes
Is the new router IP different to the old router IP?No. The new router was set up using the same ip address as the old one.
In part this was to try and avoid problems like this.
Now what I think has happened is that Oracle's VBox simply needs to be
told the new router's IP address?
I don't argue with that. However what is confusing me is that I've got
two VBox VMs (the CentOS one and a Windows10/32 one). The CentOS one
had completely fallen off the network (now fixed, see another post in
the thread), but the Windows one didn't. Additionally, I can't see
anything in the VM set up (that is from the VM's window menus, rather
than settings under /etc) that allows you to do that.
Basically, now I can reach the Internet, however, I can only access
machines on my LAN by their IP address, the local DNS lookup has stopped working.
Alternatively, has the physical IP of the machine hosting the VM's
changed? The VMs may still think the host machine still has the old IP
(that was given out by the old router) and simply need to be told what
the host machine's new IP is that is given out by new router?
All machines on my LAN are given fixed leases at the router (and that information was ported to the new router), so they should retain their
IP addresses.
Adrian
On 22/01/2023 19:22, Adrian wrote:
In message <tqjv4c$70f$1@gioia.aioe.org>, SH <i.love.spam@spam.com> writes >>> Is the new router IP different to the old router IP?
No. The new router was set up using the same ip address as the old
one. In part this was to try and avoid problems like this.
Now what I think has happened is that Oracle's VBox simply needs toI don't argue with that. However what is confusing me is that I've
be told the new router's IP address?
got two VBox VMs (the CentOS one and a Windows10/32 one). The CentOS
one had completely fallen off the network (now fixed, see another
post in the thread), but the Windows one didn't. Additionally, I
can't see anything in the VM set up (that is from the VM's window
menus, rather than settings under /etc) that allows you to do that.
Basically, now I can reach the Internet, however, I can only access >>machines on my LAN by their IP address, the local DNS lookup has
stopped working.
Ah.... you're missing a hosts file? this has FDQN (fully qualified
domain names) vs IP addresses.
Did you have a list in the old router where you could type in the FQDN
name and the machine would get the IP address from the hosts file.
For externals, this would use DNS, but for internals, that uses a hosts
file. Perhaps the new router needs a bit more info?
I run my own DNS (pi Hole) and I put my local machine names and IP
addresses in there to save me messing about creating a hosts file on
every PC and knowing that external DNS won't know about my internal
network.
Looking under "advanced settings", I find that the is a tick box for
"Cable connected". This was unticked. I've now ticked it, and I'm
back on the network, well sort of.
ip a now shows Up rather than down, and I can ping external sites,
but I can't ping any local (i.e. on my LAN) machines by name, but can
by ip address. So something somewhere means that I've lost the DNS
settings.
Am 22.01.2023 um 19:11:28 Uhr schrieb Adrian:
Looking under "advanced settings", I find that the is a tick box for
"Cable connected". This was unticked. I've now ticked it, and I'm
back on the network, well sort of.
ip a now shows Up rather than down, and I can ping external sites,
but I can't ping any local (i.e. on my LAN) machines by name, but can
by ip address. So something somewhere means that I've lost the DNS
settings.
Is the NIC in the VBox settings connected via Bridge and NOT NAT?
NAT breaks certain things here, don't use it.
Curious. The only setting that seems to work here is NAT.
If I set it to anything else, I can't ping Internet sites by name.
With NAT set I can ping Internet sites by name, and LAN sites by IP
address.
Am 23.01.2023 um 15:28:40 Uhr schrieb Adrian:
Curious. The only setting that seems to work here is NAT.
If I set it to anything else, I can't ping Internet sites by name.
With NAT set I can ping Internet sites by name, and LAN sites by IP
address.
Set it to Bridge, disconnect the virtual NIC, the connect it again.
You also need to select the physical NIC you actually use at the
computer in VBox.
https://picr.eu/images/2023/01/23/Mn7Hh.png
That's mDNS. Sounds like you lost the search-domain for ".local" as
part of the upgrade cycle. That is, the previous router had some
setting that applied the ".local" TLD to non-FQDN hostnames.
On 2023-01-22, Adrian wrote:
In message <tqk9cp$2h8ao$1@dont-email.me>, SH <i.love@spam.com> writes >>>Did you have a list in the old router where you could type in the FQDN >>>name and the machine would get the IP address from the hosts file.
The old router (and the new one, which is basically a newer version) has
the ability to put in IP addresses, names and MAC addresses for items on
the LAN, but this is fully populated. In addition to the CentOS VM,
I've got a Ubuntu PC and a stack of RPis, and they all happily work with
out any mods to the hosts file (and they basically hold the default info
for that box , references to 127.0.0.1). What has changed with the new
router is that now the various RPis no longer accept "machine_name", but
now need "machine_name.local" for ping or ssh to find the machine in
question (the Ubuntu box always did). I added another RPi to the LAN
this afternoon, and without editing anything, I was able to access it by
name straight away.
The router appears to pickup the machine name from the machine in
question, and doesn't have the ".local" bit in its version of the hosts
table.
That's mDNS. Sounds like you lost the search-domain for ".local" as
part of the upgrade cycle. That is, the previous router had some
setting that applied the ".local" TLD to non-FQDN hostnames.
I run my own DNS (pi Hole) and I put my local machine names and IP >>>addresses in there to save me messing about creating a hosts file on >>>every PC and knowing that external DNS won't know about my internal >>>network.
This is something that I've only ever done on the router. I suspect
that strictly speaking I don't need to do it there, but that is to
ensure that each machine keeps the same IP address.
DNS has nothing to do with IP address assignments to a host. It is
merely the translator of a given hostname to the machine's assigned IP >address.
In message <tqk9cp$2h8ao$1@dont-email.me>, SH <i.love@spam.com> writes
Did you have a list in the old router where you could type in the FQDN
name and the machine would get the IP address from the hosts file.
The old router (and the new one, which is basically a newer version) has
the ability to put in IP addresses, names and MAC addresses for items on
the LAN, but this is fully populated. In addition to the CentOS VM,
I've got a Ubuntu PC and a stack of RPis, and they all happily work with
out any mods to the hosts file (and they basically hold the default info
for that box , references to 127.0.0.1). What has changed with the new router is that now the various RPis no longer accept "machine_name", but
now need "machine_name.local" for ping or ssh to find the machine in
question (the Ubuntu box always did). I added another RPi to the LAN
this afternoon, and without editing anything, I was able to access it by
name straight away.
The router appears to pickup the machine name from the machine in
question, and doesn't have the ".local" bit in its version of the hosts table.
I run my own DNS (pi Hole) and I put my local machine names and IP >>addresses in there to save me messing about creating a hosts file on
every PC and knowing that external DNS won't know about my internal >>network.
This is something that I've only ever done on the router. I suspect
that strictly speaking I don't need to do it there, but that is to
ensure that each machine keeps the same IP address.
Which isn't working at all from my VM, but does from other (physical) machines on the LAN. The VM is unable to translate the name of a
machine into an IP address when it is plugged into the new router,
but with the old one, it works fine.
Am 24.01.2023 um 10:55:58 Uhr schrieb Adrian:
Which isn't working at all from my VM, but does from other (physical)
machines on the LAN. The VM is unable to translate the name of a
machine into an IP address when it is plugged into the new router,
but with the old one, it works fine.
Use Wireshark and capture for dns, LLMNR, mdns and nbns.
Check which protocol will be used.
217.169.20.20 appears in my /etc/resolv.conf as a nameserver.
If more information is needed, please advise on options to use.
In message <tqmqmn$3op39$3@dont-email.me>, Marco Moock <mo01@posteo.de> writes
Am 23.01.2023 um 15:28:40 Uhr schrieb Adrian:
Curious. The only setting that seems to work here is NAT.
If I set it to anything else, I can't ping Internet sites by name.
With NAT set I can ping Internet sites by name, and LAN sites by IP
address.
Set it to Bridge, disconnect the virtual NIC, the connect it again.
You also need to select the physical NIC you actually use at the
computer in VBox.
https://picr.eu/images/2023/01/23/Mn7Hh.png
Thanks
This is not what I'm seeing. When I open up the settings and select
network, I can see "Bridged Adapter", but the "Enable Network Adapter"
box is greyed out. No option seems to allow me to tick/un tick that
box.
With Bridged Adapter selected, under Name, I have three options :
RealtekPCIe GbE Family Controller (which is the default)
Qualcomm QCA9565 802.11b/g/n Wireless Adapter
VirtualBox Host-Only Ethernet Adapter
There is no option to enter the NIC that I'm using (enp0s3)
As an experiment last night, I swapped back to my old router, and with
that the VM worked as I expected it to, without changing any settings.
Am 24.01.2023 um 10:30:26 Uhr schrieb Dan Purgert:
That's mDNS. Sounds like you lost the search-domain for ".local" as
part of the upgrade cycle. That is, the previous router had some
setting that applied the ".local" TLD to non-FQDN hostnames.
Wrong.
.local is mDNS and will be used with IPv4/IPv6 multicast link local addresses, regardless if a router is present or not.
Although mDNS uses link-local multicast, it only works on the same
Ethernet link, NAT breaks that as I already pointed out.
Am 24.01.2023 um 13:08:04 Uhr schrieb Adrian:
217.169.20.20 appears in my /etc/resolv.conf as a nameserver.
If more information is needed, please advise on options to use.
Again, last time, I don't want to repeat myself:
Is you NIC configured as a bridge in VBox?Not currently. When I do configure it as a bridge, all ping commands
Show the output of
ip a
Did you understand that .local is mDNS and not normal Unicast DNS?Not entirely. I haven't had time (yet) to look up the definitions in
What is the content of the /etc/nsswitch.conf?
Is an mDNS client like Avahi or Bonjour running on the target machine?
A service MUST listen on 5353 UDP.
It's been a long time since I've used vbox; but as I (probably
incorrectly) recall; some of the "greyed out" things were only
accessible if you were running the vbox UI as root (or in a more
permissive vbox-something group).
As an experiment last night, I swapped back to my old router, and with
that the VM worked as I expected it to, without changing any settings.
In either case; the VM in question is getting its DNS server information
via DHCP, right? In both cases, is it the IP address of the router?
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
On 2023-01-24, Marco Moock wrote:
Am 24.01.2023 um 10:30:26 Uhr schrieb Dan Purgert:
That's mDNS. Sounds like you lost the search-domain for ".local"
as part of the upgrade cycle. That is, the previous router had
some setting that applied the ".local" TLD to non-FQDN hostnames.
Wrong.
.local is mDNS and will be used with IPv4/IPv6 multicast link local addresses, regardless if a router is present or not.
Although mDNS uses link-local multicast, it only works on the same
Ethernet link, NAT breaks that as I already pointed out.
You seem to have misread the interaction.
Old router -> unqualified hostnames (i.e. "rpi3") resolved properly to
hosts on the LAN.
New router -> unqualified hostnames (i.e "rpi3") do not resolve; BUT
their mDNS hostnames (i.e. "rpi3.local") do.
Therefore -> "old router" had been providing the appropriate mDNS search-domain (".local").
I'd imagine this to be most likely provided
as part of the DHCP options; although it could potentially have been
part of the router's own internal DNS server configuration.
This is not what I'm seeing. When I open up the settings and select
network, I can see "Bridged Adapter", but the "Enable Network
Adapter" box is greyed out. No option seems to allow me to tick/un
tick that box.
RealtekPCIe GbE Family Controller (which is the default)
Qualcomm QCA9565 802.11b/g/n Wireless Adapter
VirtualBox Host-Only Ethernet Adapter
There is no option to enter the NIC that I'm using (enp0s3)
With Bridged Adapter selected, under Name, I have three options :
RealtekPCIe GbE Family Controller (which is the default)
Qualcomm QCA9565 802.11b/g/n Wireless Adapter
VirtualBox Host-Only Ethernet Adapter
There is no option to enter the NIC that I'm using (enp0s3)
The config
options for the VM wrapper (for want of a better name) don't seem to
be as extensive with VMWare, but the networking shows up as NAT.
Network connection options are :
Bridged
Nat (selected)
Host-only
Custom
Lan segment
Am I right in drawing the conclusion that the problem isn't with the
VM itself, but with the VM hosting software and the way it talks to
the new router ?
Adrian wrote:
With Bridged Adapter selected, under Name, I have three options :
RealtekPCIe GbE Family Controller (which is the default)
Qualcomm QCA9565 802.11b/g/n Wireless Adapter
VirtualBox Host-Only Ethernet Adapter
those look like windows style device names, presumably that's your host O/S?
There is no option to enter the NIC that I'm using (enp0s3)
that looks like linux style device nsames, I presume thsat's your guest
O/S, the names won't be the same, the guest won't really know what
actual NICs the host has.
All looks relatively sane though ...
Am 24.01.2023 um 10:35:50 Uhr schrieb Adrian:
This is not what I'm seeing. When I open up the settings and select
network, I can see "Bridged Adapter", but the "Enable Network
Adapter" box is greyed out. No option seems to allow me to tick/un
tick that box.
Shut down the virtual machine. Enabling the NIC in VBox is like
connecting a PCI card to a physical system.
Don't confuse "Enable" with "Connected".
RealtekPCIe GbE Family Controller (which is the default)
Qualcomm QCA9565 802.11b/g/n Wireless Adapter
VirtualBox Host-Only Ethernet Adapter
There is no option to enter the NIC that I'm using (enp0s3)
en is Ethernet --> select the Realtek NIC.
I then pinged one of my Pis, and it took 23 seconds to tell me "Name
or service not known"
I also pinged an external site, and again, it took 23 seconds to tell
me "Name or service not known"
$ cat /etc/resolv.conf
# Generated by NetworkManager
search localdomain
nameserver 217.169.20.20
nameserver 217.169.20.21
Am 24.01.2023 um 18:29:57 Uhr schrieb Adrian:
I then pinged one of my Pis, and it took 23 seconds to tell me "Name
or service not known"
I also pinged an external site, and again, it took 23 seconds to tell
me "Name or service not known"
Please show the output of
cat /etc/resolv.conf
resolvectl
Am 26.01.2023 schrieb Adrian <bulleid@ku.gro.lioff>:
$ cat /etc/resolv.conf
# Generated by NetworkManager
search localdomain
nameserver 217.169.20.20
nameserver 217.169.20.21
Please also
cat /etc/nsswitch.conf
If you want to use .local, is avahi installed and running?
hosts: files dns myhostname
avahi is not installed. I am not bothered whether I use name or
name.local to ping / ssh, but I would like to be able to use at least
one of them.
Am 26.01.2023 schrieb Adrian <bulleid@ku.gro.lioff>:
hosts: files dns myhostname
no mDNS enabled, mDNS lookup (default for .local) WILL NOT WORK.
avahi is not installed. I am not bothered whether I use name or
name.local to ping / ssh, but I would like to be able to use at least
one of them.
.local is intended for mDNS, so install avahi to be able to resolve
these names properly.
systemctl status shows avahi-daemon running.
However, I still don't see any difference in the nsswitch.conf file.
Am 26.01.2023 schrieb Adrian <bulleid@ku.gro.lioff>:
systemctl status shows avahi-daemon running.
However, I still don't see any difference in the nsswitch.conf file.
Use Wireshark/other network sniffer and filter for mDNS.
Do you see any queries?
Am 27.01.2023 um 15:54:08 Uhr schrieb Adrian:
That appears to be going straight out to the Internet to do a lookup,
so it is hardly surprising that I can't get a result. Why doesn't it
check with the router for local names ?
Check nsswitch.conf.
mdns must be present here for hosts.
See here for examples: >https://atwillys.de/content/linux/mdns-on-debian-ubuntu-installation/
That appears to be going straight out to the Internet to do a lookup,
so it is hardly surprising that I can't get a result. Why doesn't it
check with the router for local names ?
Why doesn't it check with the router for local names ?
Adrian wrote:
Why doesn't it check with the router for local names ?
Because not all routers provide local DNS
hosts: files dns myhostname
One works fine (seemingly without mdns), the other one doesn't.
Am 27.01.2023 um 23:50:41 Uhr schrieb Adrian:
One works fine (seemingly without mdns), the other one doesn't.
Does it query via mDNS?
and I've not been able to work out how I can see the output, so I'm
unable to see what it is doing. Whereas on the "non-working" VM
tshark gives me output (as posted up thread).
Am 28.01.2023 um 13:49:29 Uhr schrieb Adrian:
and I've not been able to work out how I can see the output, so I'm
unable to see what it is doing. Whereas on the "non-working" VM
tshark gives me output (as posted up thread).
Do you have any filters on tshark?
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 299 |
| Nodes: | 16 (2 / 14) |
| Uptime: | 72:51:57 |
| Calls: | 6,694 |
| Calls today: | 4 |
| Files: | 12,228 |
| Messages: | 5,346,772 |
| Posted today: | 1 |
