[opensuse] router DHCP suddenly not talking to one machine (10.1)
One of my desktop machines is dual-boot and lived happily attached to my SMC router, which has a built in DHCP server. After a change of network card the DHCP server saw and assigned an address to it, as usual. After booting into Windows (network working again fine) I went back to Suse 10.1 but there was no IP address. The DHCP client keeps waiting. The DHCP router does not acknowledge the presence of this network card in the status window, even though all lights are green (and connection works in Windows). I reconfigured the card a few times and rebooted the router. I disabled the MAC address control on the router and the other machines are fine, all acknowledged by the router and given the usual address. Any clues? Thanks, Carlos FL -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Carlos F Lange wrote:
One of my desktop machines is dual-boot and lived happily attached to my SMC router, which has a built in DHCP server. After a change of network card the DHCP server saw and assigned an address to it, as usual. After booting into Windows (network working again fine) I went back to Suse 10.1 but there was no IP address. The DHCP client keeps waiting. The DHCP router does not acknowledge the presence of this network card in the status window, even though all lights are green (and connection works in Windows).
I reconfigured the card a few times and rebooted the router. I disabled the MAC address control on the router and the other machines are fine, all acknowledged by the router and given the usual address.
Any clues?
Thanks, Carlos FL
Some thoughts.... a) Does the NIC communicate with a static IP address from SuSE? b) Which NIC? c) Has the network information been reconfigured on SuSE? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFGZROBasN0sSnLmgIRAro3AJ4pUa4JZ0iYuF0dTszuSXPxC0IhqQCaA/b8 3xDrpIv2nRs17y/6L95pG+0= =HO/c -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tue June 5 2007 01:40, G T Smith wrote:
Carlos F Lange wrote:
One of my desktop machines is dual-boot and lived happily attached to my SMC router, which has a built in DHCP server. After a change of network card the DHCP server saw and assigned an address to it, as usual. After booting into Windows (network working again fine) I went back to Suse 10.1 but there was no IP address. The DHCP client keeps waiting. The DHCP router does not acknowledge the presence of this network card in the status window, even though all lights are green (and connection works in Windows).
I reconfigured the card a few times and rebooted the router. I disabled the MAC address control on the router and the other machines are fine, all acknowledged by the router and given the usual address.
Any clues?
Thanks, Carlos FL
Some thoughts....
a) Does the NIC communicate with a static IP address from SuSE?
No. Even an attempt to ping the router (192.168.2.1) gives error: "Connect: Network is unreachable." "rcnetwork status" gives "eth0 dhcpd is still waiting for data."
b) Which NIC?
nVidia Corporation MCP51 Ethernet Controller
c) Has the network information been reconfigured on SuSE?
I deleted the configuration in the "Network Card" configuration in Yast and reconfigured it (Traditional Method with ifup), but there was no change. change. DHCP Options are default (unset apart from Hostname AUTO) and the same settings are still working in the other machines. Carlos FL -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Carlos F Lange wrote:
On Tue June 5 2007 01:40, G T Smith wrote:
Carlos F Lange wrote:
I reconfigured the card a few times and rebooted the router. I disabled the MAC address control on the router and the other machines are fine, all acknowledged by the router and given the usual address.
Any clues?
Thanks, Carlos FL Some thoughts....
a) Does the NIC communicate with a static IP address from SuSE?
No. Even an attempt to ping the router (192.168.2.1) gives error: "Connect: Network is unreachable." "rcnetwork status" gives "eth0 dhcpd is still waiting for data."
I meant with a static address assigned to the NIC the above means that you are still attempting to obtain an address by DHCP... This would establish whether the driver had an issue, or something odd has happened to the DHCP client. A very faint possibility is that there may be an issue with connection negotiation (Duplex, 10/100 mbps etc).
b) Which NIC?
nVidia Corporation MCP51 Ethernet Controller
Is this the same NIC as on your other machines. If you think is so check chipset on card... NIC card manufacturers have a nasty habit of revising the chipset under the hood. <snip>
Carlos FL
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFGZYNLasN0sSnLmgIRAi48AJ9q5r933Y/0p2iLG1c5ETnSZ4dUMwCffZox /JCHaKUBrs/vappw28K1EUw= =oQfs -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tue June 5 2007 09:37, G T Smith wrote:
a) Does the NIC communicate with a static IP address from SuSE?
No. Even an attempt to ping the router (192.168.2.1) gives error: "Connect: Network is unreachable." "rcnetwork status" gives "eth0 dhcpd is still waiting for data."
I meant with a static address assigned to the NIC the above means that you are still attempting to obtain an address by DHCP... This would establish whether the driver had an issue, or something odd has happened to the DHCP client.
I did now assign the usual IP statically to the card in Yast. "ifconfig" shows the address, but I cannot ping anything in the LAN.
A very faint possibility is that there may be an issue with connection negotiation (Duplex, 10/100 mbps etc).
Would there be any log for this negotiation?
b) Which NIC?
nVidia Corporation MCP51 Ethernet Controller
Is this the same NIC as on your other machines. If you think is so check chipset on card... NIC card manufacturers have a nasty habit of revising the chipset under the hood.
No. Every machine is unique. This one just got a new NIC, which worked fine in Suse for 2 days before I first booted into Windows. When I booted back into Linux, then my problems started. That is why I suspected the DHCP server had some hiccup, but resetting it did not help. Carlos FL -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tuesday 05 June 2007, Carlos F Lange wrote: [...]
b) Which NIC?
nVidia Corporation MCP51 Ethernet Controller
Is this the same NIC as on your other machines. If you think is so check chipset on card... NIC card manufacturers have a nasty habit of revising the chipset under the hood.
No. Every machine is unique. This one just got a new NIC, which worked fine in Suse for 2 days before I first booted into Windows. When I booted back into Linux, then my problems started. That is why I suspected the DHCP server had some hiccup, but resetting it did not help.
Carlos FL
---------------------- I'm going to ask a silly question and make a silly suggestion. When you say you booted into Windows then booted into Linux later, did you shutdown the machine first or just reboot it selecting Linux? Next, if you did the above, try shutting down the computer, as in turn off, then booting directly into Linux when you power back up to see if it comes back online. Lee -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tue June 5 2007 12:06, BandiPat wrote:
I'm going to ask a silly question and make a silly suggestion. When you say you booted into Windows then booted into Linux later, did you shutdown the machine first or just reboot it selecting Linux?
Next, if you did the above, try shutting down the computer, as in turn off, then booting directly into Linux when you power back up to see if it comes back online.
Not so silly, I've had that too. The answer is yes, I did shutdown the machine and wait. CFL -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tuesday 05 June 2007, Carlos F Lange wrote:
On Tue June 5 2007 12:06, BandiPat wrote:
I'm going to ask a silly question and make a silly suggestion. When you say you booted into Windows then booted into Linux later, did you shutdown the machine first or just reboot it selecting Linux?
Next, if you did the above, try shutting down the computer, as in turn off, then booting directly into Linux when you power back up to see if it comes back online.
Not so silly, I've had that too. The answer is yes, I did shutdown the machine and wait.
CFL
Is this a plug and play capable nic? I wonder if this is one of those cards where it might get an interrupt and port assignment set by the plug and play software in windows, and then be unable to use those settings in Linux (for whatever reason). Lots of nics used to be set up this way a few years back and some were so cranky they came with their own configuration software disk, and god help you if you lost it. I would check your bios to see if there is a setting for Plug and Play OS installed, and tell it NO (or tell it the opposite of what it is currently set to). The NO choice tells the bios to do the plug and play configuration rather than leaving it up to the OS. -- _____________________________________ John Andersen -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Carlos F Lange wrote:
On Tue June 5 2007 09:37, G T Smith wrote:
<snip>
A very faint possibility is that there may be an issue with connection negotiation (Duplex, 10/100 mbps etc).
Would there be any log for this negotiation?
What are the reported packet stats from ifstatus/netstat like? Lots of incoming error packets would be a strong indicator. With a consumer router you are unlikely to get this kind of report of connection status. So even if the driver configuration switches were available to force the type of connection it would be guesswork to get the right configuration in place. The suggestion made elsewhere of an iffy cable is worth considering, but could be easily eliminated by connecting with a cable you know has been working before.
b) Which NIC? nVidia Corporation MCP51 Ethernet Controller
<snip>
No. Every machine is unique. This one just got a new NIC, which worked fine in Suse for 2 days before I first booted into Windows. When I booted back into Linux, then my problems started. That is why I suspected the DHCP server had some hiccup, but resetting it did not help.
The suggestion of making certain that all traces of the previous card configuration are removed which has also been made is also sound. I do not entirely trust YaST in this area. If the machine had not powered down for 2 days before booting into windows I would take a hard look at the udev persistant names rules and associated settings. These can generate the odd surprise.
Carlos FL
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFGZcNpasN0sSnLmgIRAtE/AJ9vPwZEzUnSoMt5yI357K4emD9lswCgr0x7 /BaUVitbAO3OPxnW2uuVxog= =4YxZ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
I solved the problem with brute force (see my other e-mail today), but I wanted to reply here to see if learn some more. On Tue June 5 2007 14:11, G T Smith wrote:
Carlos F Lange wrote:
On Tue June 5 2007 09:37, G T Smith wrote:
<snip>
A very faint possibility is that there may be an issue with connection negotiation (Duplex, 10/100 mbps etc).
Would there be any log for this negotiation?
What are the reported packet stats from ifstatus/netstat like? Lots of incoming error packets would be a strong indicator.
ifstatus:
eth0 device: nVidia Corporation MCP51 Ethernet Controller (rev a1)
eth0 configuration: eth-id-00:15:f2:bf:d8:2c
eth0 dhcpcd is still waiting for data
eth0 is up
2: eth0:
With a consumer router you are unlikely to get this kind of report of connection status. So even if the driver configuration switches were available to force the type of connection it would be guesswork to get the right configuration in place.
The suggestion made elsewhere of an iffy cable is worth considering, but could be easily eliminated by connecting with a cable you know has been working before.
Not the case.
b) Which NIC?
nVidia Corporation MCP51 Ethernet Controller
<snip>
No. Every machine is unique. This one just got a new NIC, which worked fine in Suse for 2 days before I first booted into Windows. When I booted back into Linux, then my problems started. That is why I suspected the DHCP server had some hiccup, but resetting it did not help.
The suggestion of making certain that all traces of the previous card configuration are removed which has also been made is also sound. I do not entirely trust YaST in this area.
I tried deleting the content in: /etc/resolv.conf /etc/HOSTNAME /etc/sysconfig/network/routes
If the machine had not powered down for 2 days before booting into windows I would take a hard look at the udev persistant names rules and associated settings. These can generate the odd surprise.
OK. I deleted the 2 entries in /etc/udev/rules.d/30-net_persistent_names.rules But this cleaning did not solve the problem. Did I forget anything? Thanks for all the help, Carlos FL -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Carlos F Lange wrote:
I solved the problem with brute force (see my other e-mail today), but I wanted to reply here to see if learn some more.
On Tue June 5 2007 14:11, G T Smith wrote:
Carlos F Lange wrote:
On Tue June 5 2007 09:37, G T Smith wrote: <snip>
A very faint possibility is that there may be an issue with connection negotiation (Duplex, 10/100 mbps etc). Would there be any log for this negotiation? What are the reported packet stats from ifstatus/netstat like? Lots of incoming error packets would be a strong indicator.
ifstatus: eth0 device: nVidia Corporation MCP51 Ethernet Controller (rev a1) eth0 configuration: eth-id-00:15:f2:bf:d8:2c eth0 dhcpcd is still waiting for data eth0 is up 2: eth0:
mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 00:15:f2:bf:d8:2c brd ff:ff:ff:ff:ff:ff inet6 fe80::215:f2ff:febf:d82c/64 scope link valid_lft forever preferred_lft forever netstat is too big, but it did not have _anything_ under "Active Internet connections".
Netstat would be without switches. Should have been a bit more specific...... I should also have suggested ifconfig NOT ifstatus.... my mistake... <snip>
I tried deleting the content in: /etc/resolv.conf /etc/HOSTNAME /etc/sysconfig/network/routes
What we are more interested in is hardware configuration related files. These are usually found in /etc/sysconfig/[network|hardware] directories. These three should change on connection anyway.
If the machine had not powered down for 2 days before booting into windows I would take a hard look at the udev persistant names rules and associated settings. These can generate the odd surprise.
OK. I deleted the 2 entries in /etc/udev/rules.d/30-net_persistent_names.rules
But this cleaning did not solve the problem. Did I forget anything?
What was not clear in the original Post was that this was an on board chipset, these can require specially tailored versions of the drivers that are usually provided by the MOBO manufacturer. (I have also had problems with Windows and Linux in the past with generic drivers and onboard NICs). This rather sounds like it should have been a case of using NDISWRAPPER with the windows drivers, or checking whether the MOBO manufacturer or nvidia has more recent linux drivers. If you installed the MOBO drivers under windows (or they were pre-installed) they have may have made firmware changes which were incompatible with your current Linux drivers. BTW Initial NDISWRAPPER based configuration is usually best done outside of YaST. Then you can use YaST to setup the appropriate NDISWRAPPER based configuration.
Thanks for all the help, Carlos FL
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFGZ8UEasN0sSnLmgIRAqKEAKC1ZYLXlmm3C2Z0aFRTS+/8veToHACgtSRC CW1sZ44HAzDG/hopGS/BmYo= =vH3Z -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2007-06-05 at 11:49 -0600, Carlos F Lange wrote:
A very faint possibility is that there may be an issue with connection negotiation (Duplex, 10/100 mbps etc).
Would there be any log for this negotiation?
No, but there are two tools that can check the card: ethtool (try "ethtool eth0" first), and mii-diag - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFGZd34tTMYHG2NR9URAuYRAJ93YM9WF8WsyjYR/rp/Jc8gyyE5aQCghNgl KRK00YrNMrf6TrfPsy2SqRI= =9yKe -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (5)
-
BandiPat
-
Carlos E. R.
-
Carlos F Lange
-
G T Smith
-
John Andersen