Network stopped working

[d_rylndr]

I no longer get IPs from my B3. The connections, both wired and wireless is seen but I never get any IP from it. The icon in my network manager just spins trying to connect.
It suddenly stopped working after I mounted the b3s shares on a new workstation installation. The fstab is the same I use on all my machines and have been using for many years. This should in no way break b3's network interface...

The B3 still could be reached from internet though it's easyfind adress so it is up and running, just not the network...

This happened once before, I then solved it by reinstalling the b3. Tried that now but it didn't help this time. I even tried several times.Now I can't reach b3 from the easyfind address either...

As I did a software system reinstallation my only conclusion is hardware malfuntion. Can I replace the network card in the b3? Is this a standard part I can get anywhere or do I need to get it from excito?

[Ubi]

You are being very cryptic.
* You can reach the B3 but you claim the network is down. Unless you soldered a serial interface to the B3, this is impossible.
* What I think you mean is that you can reach the B3 over the WAN port, but not the LAN port. If so, it means you have access to log files, you can fiddle with the firewall and even sniff network traffic over the NIC.
* You also claim the the connections are "seen" but you do not get a DHCP adrress, yet you blame this on hardware.
* You have problems connection to WLAN, yet you blame this on a faulty NIC, which is a completely different IC on the board.
* You suggest the problem lies in DHCP, but you think this is related to fstab, which regulates disk
mounting parameters. However, do did not seem to have tested IP connectivity directly, or ran a port scan.
* You see a loss of connectivity to the external nic *directly* after you alter software, but yet you blame it on hardware.
* The fact that you tried to connect on the easyfind address rather than the external IP address directly suggests to me you are not an expert on IP and network connectivity, yet you jump to conclusions after a few tests and immediately format the disk.

But to answer your question: No, the NIC is not user-replaceable. However I do not think the hardware is the problem.

[d_rylndr]

You are being very cryptic.

You think so? Sorry for that. I'll clarify.

* You can reach the B3 but you claim the network is down. Unless you soldered a serial interface to the B3, this is impossible.

Yes. Can not reach from LAN, b3 is seen in network manager, but I can not connect. Is not given any IP. I could, before I reinstalled b3, now that is no longer possible, connect to the b3 over the internet through it's easyfind-address.

* What I think you mean is that you can reach the B3 over the WAN port, but not the LAN port. If so, it means you have access to log files, you can fiddle with the firewall and even sniff network traffic over the NIC.

Yes. But no longer. After this happened I replaced the b3 with an old router so I can at least have internet access. I connected the b3 to the router as well (b3 WAN > router LAN) but the b3 does not show up as an attached device in the router so I can not ssh into it using its IP no longer.

* You also claim the the connections are "seen" but you do not get a DHCP adrress, yet you blame this on hardware.

Yes, seen in the network managers of all clients in my network, none of which is given an IP from the b3 who acts as a router+file server.
I guess, not suggests, hardware malfunction as I did a reinstallation of the b3 system, resetting it to factory settings which didn't help. If it was some software fault don't you think it would be solved by reinstalling?

* You have problems connection to WLAN, yet you blame this on a faulty NIC, which is a completely different IC on the board.

Yes, forgive me for not being a network technician by profession but a mere layman. I am probably wrong on this point.

* You suggest the problem lies in DHCP, but you think this is related to fstab, which regulates disk
mounting parameters.

No, I don'ty think it is fstab related, I just reported what I did just before the b3 malfuntioned.

However, do did not seem to have tested IP connectivity directly, or ran a port scan.

Please tell me how and I would be glad to do it.

* You see a loss of connectivity to the external nic *directly* after you alter software, but yet you blame it on hardware.

I didn't alter software, I mounted the b3's shared disks on a client machine after which the b3 stopped working. The only factor that is worth noting here is that it was a new installation on the client machine so it was the first time it mounted the b3's disks. The fstab the client machine uses however is an old one I keep and use for all my client machines.

* The fact that you tried to connect on the easyfind address rather than the external IP address directly suggests to me you are not an expert on IP and network connectivity, yet you jump to conclusions after a few tests and immediately format the disk.

Good catch Sherlock! Yes, I reinstalled as it was what got the machine working last time.

But to answer your question: No, the NIC is not user-replaceable. However I do not think the hardware is the problem.

Alright. What would you guess be?

thank you.

[Ubi]

The thing is this; you may have had a simple case of the DHCP server borking, which you could have dignosed easily by going into the WAN port and looking at the logs. Alas.

Now that you formatted the disk, your current problem could be completely different. All that remains is that you cannot access a freshly installed B3 over the network. This could be because your network is expecting a different set of configurations than the default install is giving. Or something else.

AFAIK the default install sets itself up as a DHCP client on the WAN port. You did not describe if there is an upstream router on the WAN port, or even how your B3 is connected to the network. This is essential information to be able to help you resolving your problems. Please describe this.

[Gordon]

Actually, if I remember correctly the default install will accept DHCP offerings on both LAN and WAN and fall back to 192.168.10.1 on the LAN side if there is no active DHCP server. So essentially all you need to do is boot the machine without any network cable attached to it and then plug a single PC straight to the LAN port. At that point you may or may not receive an IP address, based on whether the B3 started the DHCP server or not, and of course if you don't you'll need to set a manual address on that PC.

Obviously having reinstalled the B3 there's no chance of checking logs. A puzzling statement being made here is that prior to the reinstall the machine could still be accessed through its easyfind address, which implies that TS has some (other) way to get onto the internet. I agree with Ubi that at this point there is no solid reason to suspect hardware failure - lots of stuff may have gone wrong here in the process of unplugging, reinstalling, reconnecting. It could have gotten a different WAN IP for what we know. That the B3 does not register in the router can also be an issue with the router firmware - sometimes DHCP servers simply fail to retrieve the name of the connected device and depending on the firmware they may either not show these at all in the management interface or not be able to distinguish between them and only show one of them (typically the printed name will be '*')

[Ubi]

To make things increasingly complex, the OP did not state if there was an upstream router that may actually have been the acting DHCP server on the network. This means the original problem was not in the B3 at all. Without the network topology this is anyones guess

[d_rylndr]

...So essentially all you need to do is boot the machine without any network cable attached to it and then plug a single PC straight to the LAN port. At that point you may or may not receive an IP address, based on whether the B3 started the DHCP server or not, and of course if you don't you'll need to set a manual address on that PC...

Thank you. I did try that now, rebooted without any network cables connected. Plugged one in from a client machine to b3 LAN. - Nothing happens at all.

Obviously having reinstalled the B3 there's no chance of checking logs. A puzzling statement being made here is that prior to the reinstall the machine could still be accessed through its easyfind address, which implies that TS has some (other) way to get onto the internet.

Yes, unfortunately. But given that I've had a similar breakage before at which point a reinstallation made everything working again. I don't have much custom settings and installations so instead of spending hours again trying to solve the issue I took the easy road and just did a reinstallation. Only it turned out not being the easy road this time.
Here is the forum post from that previous failure: http://forum.excito.net/viewtopic.php?f=9&t=3521
And yes, I have internet in my telephone. With that I could reach the b3.

To make things increasingly complex, the OP did not state if there was an upstream router that may actually have been the acting DHCP server on the network. This means the original problem was not in the B3 at all. Without the network topology this is anyones guess

I did not because I had none. I replaced the b3 router with an old netgear router after the failure so I could have internet access at all.

Ubi, do you have an checkpoint form with all information you need from me? I am not a network technician. I try to give as much information as I can but clearly I don't think of everything. You keep saying I should provide more information. Please tell me exactly what you need.

[Ubi]

The checkpoint list is a very good suggestion actually, but unfortunately we have none. What I'm interested in is how your external internet is connected to your B3. What I think is that you have a UTP cable coming out of a wall that you stick directly into the WAN port of the B3, and you either get DHCP from that cable, or you are required to enter a fixed IP address provided by your internet provider. Then, you have a cable from the LAN port of the B3 going into a switch. From that switch theres cables going to other computers. Is this accurate?

[d_rylndr]

The checkpoint list is a very good suggestion actually, but unfortunately we have none. What I'm interested in is how your external internet is connected to your B3. What I think is that you have a UTP cable coming out of a wall that you stick directly into the WAN port of the B3, and you either get DHCP from that cable, or you are required to enter a fixed IP address provided by your internet provider. Then, you have a cable from the LAN port of the B3 going into a switch. From that switch theres cables going to other computers. Is this accurate?

Yes exactly my setup. I have no fixed IP from my provider.

[Ubi]

So you get a dynamic IP from the cable into the B3. What if you remove the B3, attach the switch or hub (not a router!) directly to that cable coming out of the wall and attach a couple computers to that switch. Are they all given a unique IP address?

[d_rylndr]

No, they're not. Only one machine in the hub gets an ip.

[Ubi]

So how does the upstream know which machine to provide the IP to?

(the thing is, if this process acts like a DHCP, it may trick the B3 into thinking there is already a DHCP on the network. This makes the B3 disable its own DHCP server, resulting in the original problem you perceived)

[Gordon]

...So essentially all you need to do is boot the machine without any network cable attached to it and then plug a single PC straight to the LAN port. At that point you may or may not receive an IP address, based on whether the B3 started the DHCP server or not, and of course if you don't you'll need to set a manual address on that PC...

Thank you. I did try that now, rebooted without any network cables connected. Plugged one in from a client machine to b3 LAN. - Nothing happens at all.

By saying nothing happens at all I conclude that you did not set a manual IP address on the client machine. Is that correct?

Right then. Let's try something else. Shut down the B3 and attach the LAN interface to your router. Now boot and wait for the blue light to appear. Next on your desktop open a terminal and run `ip addr`. This should give you the current network address that is assigned to your machine (most likely bound to eth0). Let's assume it says 192.168.1.6/24

Now run the following command

Code: Select all

for ip in 192.168.1.{1..254}; do ping -c 1 -t 1 -w 1 $ip > /dev/null && echo "${ip} is up"; done

Change the first three numbers (i.e. 192.168.1) in there to match those of the IP of the desktop machine.

This will give you a list of all active IP addresses on your LAN and one of them should be the B3

[d_rylndr]

Thanks.

the b3 doesn't get an IP doing that.
It does hovever get one if I plug in a rescue system usb stick (http://wiki.excito.org/wiki/index.php/Rescue_System) and reboot into that.
I ssh into the rescue system.
I mount /dev/sda1 at /mnt

I very carefully selected not to format or partition when I reinstalled the system before, yet /mnt/home does not contain any of the user homes (only default admin/storage/web) and /mnt/home/admin is empty except for the default installation files.
/home shouldn't be erased reinstalling without partition or format!?

df -h reports sda1 to be a 9.8G system, but doesn't list sda2 or sda3 at all.

Code: Select all

Filesystem                Size      Used Available Use% Mounted on
/dev/root                 6.0M      3.9M      1.8M  68% /
tmpfs                   252.1M     44.0K    252.0M   0% /tmp
df: /mnt/usb: No such file or directory
/dev/sda1                 9.8G    942.9M      8.4G  10% /mnt

while parted -l reports all partitions of the 2Tb disk.

Code: Select all

Model: ATA WDC WD20EARS-00M (scsi)
Disk /dev/sda: 2000GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End     Size    File system     Name  Flags
 1      17.4kB  10.7GB  10.7GB  ext3            root
 2      10.7GB  1999GB  1989GB                  home
 3      1999GB  2000GB  1074MB  linux-swap(v1)  swap

It doesn't report the file system type of sda2, why!?
I'm not allowed to mount sda2

Code: Select all

mount: mounting /dev/sda2 on /tmp/sda2mnt failed: Invalid argument

which is my storage disk...

/mnt/home/storage doesn't contain anything except the default folders.

So I now have two questions:
1. why can't I mount my sda2
2. How can I get a working system back when reinstalling doesn't seem to help? I do NOT want to partition the whole disk. Of course I have a backup of the file server storage disk (sda2) but it is a hassle to restore almost 2Tb of data...

[Ubi]

The reason is that the rescue disk cannot handle lvm, which indeed is kinda annoying. Your data is most likely ok.

Anyway, the fun bit is that you just validated the hardware.