I have had an issue with F18 NetworkManager (yes I know) since installing. It does not seem to get on with my Netgear DG834v2 wired router/ADSL modem. F17 behaves perfectly well. Both installations are stock Gnome installs and I have done no configuration to the network and there is only this one machine connected to the router..
The network goes AWOL for several minutes and then returns. This is particularly irritating while internet browsing but also affects yum updates when one gets frequent '[Errno 14] curl#6 - "Could not resolve host' errors.
I have monitored the the ethernet interface in non-promiscuous mode with wireshark with just NTP accessing the network and discovered that three ARP requests are sent for who has 192.168.0.1 separated by one second and, if there is no response, three ARP broadcasts are made for who has 192.168.0.1 and if there is still no response the three broadcasts are repeated every 60 seconds.
Sometimes it waits for 120 seconds then sends broadcasts without first addressing the router.
The following is a snippet of the wireshark output
During this time the network is unavailable to applications.
In particular I cannot log into the router to see if the connection to my ISP has been lost. When I do finally get to the router, it reports it has been up continuously and there is nothing in the logs.
I can prevent the network loss by logging into the router via the web interface and running the Router Statistics monitor which updates through the browser every 5 seconds.
Does anyone know the inner workings of NetworkManager and can suggest any configuration changes to make it behave more like F17 and stop pestering the router?
F18 Packages
kernel-PAE-3.8.3-203.fc18.i686
NetworkManager-0.9.8.0-1.fc18.i686
dbus-1.6.8-2.fc18.i686
F17 Packages
kernel-PAE-3.8.3-103.fc17.i686
NetworkManager-0.9.6.4-3.fc17.i686
dbus-1.4.10-7.fc17.i686
From lspci :-
00:0b.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ (rev 10)
From /var/log/messages :-
kernel: [ 18.876141] 8139cp: 8139cp: 10/100 PCI Ethernet driver v1.3 (Mar 22, 2004)
kernel: [ 18.879198] 8139cp 0000:00:0b.0: This (id 10ec:8139 rev 10) is not an 8139C+ compatible chip, use 8139too
kernel: [ 19.219403] 8139too: 8139too Fast Ethernet driver 0.9.28
kernel: [ 19.296291] ACPI: PCI Interrupt Link [LNKB] enabled at IRQ 10
kernel: [ 19.374273] 8139too 0000:00:0b.0 eth0: RealTek RTL8139 at 0xe883a000, 00:14:6c:81:aa:63, IRQ 10
---------- Post added at 03:37 PM ---------- Previous post was at 11:17 AM ----------
I posted too soon. I cannot prevent the network loss by running the Router Statistics monitor after all.
---------- Post added at 07:28 PM ---------- Previous post was at 03:37 PM ----------
Well I now have one instance of the problem under F17, so maybe it is a flaky router???
The network goes AWOL for several minutes and then returns. This is particularly irritating while internet browsing but also affects yum updates when one gets frequent '[Errno 14] curl#6 - "Could not resolve host' errors.
I have monitored the the ethernet interface in non-promiscuous mode with wireshark with just NTP accessing the network and discovered that three ARP requests are sent for who has 192.168.0.1 separated by one second and, if there is no response, three ARP broadcasts are made for who has 192.168.0.1 and if there is still no response the three broadcasts are repeated every 60 seconds.
Sometimes it waits for 120 seconds then sends broadcasts without first addressing the router.
The following is a snippet of the wireshark output
Code:
70 2040.223887000 Netgear_81:aa:63 Netgear_68:e2:24 ARP 42 192.168.0.2 is at 00:14:6c:81:aa:63
71 2083.724084000 Netgear_68:e2:24 Netgear_81:aa:63 ARP 60 Who has 192.168.0.2? Tell 192.168.0.1
72 2083.724167000 Netgear_81:aa:63 Netgear_68:e2:24 ARP 42 192.168.0.2 is at 00:14:6c:81:aa:63
73 2178.472669000 Netgear_81:aa:63 Netgear_68:e2:24 ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
74 2179.474684000 Netgear_81:aa:63 Netgear_68:e2:24 ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
75 2180.476690000 Netgear_81:aa:63 Netgear_68:e2:24 ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
76 2195.828201000 Netgear_81:aa:63 Broadcast ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
77 2196.828785000 Netgear_81:aa:63 Broadcast ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
78 2197.830705000 Netgear_81:aa:63 Broadcast ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
79 2198.833013000 Netgear_81:aa:63 Broadcast ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
...
215 2362.022701000 Netgear_81:aa:63 Broadcast ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
216 2363.024708000 Netgear_81:aa:63 Broadcast ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
217 2406.762783000 Netgear_81:aa:63 Broadcast ARP 42 Who has 192.168.0.1? Tell 192.168.0.2
218 2406.763199000 Netgear_68:e2:24 Netgear_81:aa:63 ARP 60 192.168.0.1 is at 00:14:6c:68:e2:24
219 2585.076950000 Netgear_68:e2:24 Netgear_81:aa:63 ARP 60 Who has 192.168.0.2? Tell 192.168.0.1In particular I cannot log into the router to see if the connection to my ISP has been lost. When I do finally get to the router, it reports it has been up continuously and there is nothing in the logs.
I can prevent the network loss by logging into the router via the web interface and running the Router Statistics monitor which updates through the browser every 5 seconds.
Does anyone know the inner workings of NetworkManager and can suggest any configuration changes to make it behave more like F17 and stop pestering the router?
F18 Packages
kernel-PAE-3.8.3-203.fc18.i686
NetworkManager-0.9.8.0-1.fc18.i686
dbus-1.6.8-2.fc18.i686
F17 Packages
kernel-PAE-3.8.3-103.fc17.i686
NetworkManager-0.9.6.4-3.fc17.i686
dbus-1.4.10-7.fc17.i686
From lspci :-
00:0b.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ (rev 10)
From /var/log/messages :-
kernel: [ 18.876141] 8139cp: 8139cp: 10/100 PCI Ethernet driver v1.3 (Mar 22, 2004)
kernel: [ 18.879198] 8139cp 0000:00:0b.0: This (id 10ec:8139 rev 10) is not an 8139C+ compatible chip, use 8139too
kernel: [ 19.219403] 8139too: 8139too Fast Ethernet driver 0.9.28
kernel: [ 19.296291] ACPI: PCI Interrupt Link [LNKB] enabled at IRQ 10
kernel: [ 19.374273] 8139too 0000:00:0b.0 eth0: RealTek RTL8139 at 0xe883a000, 00:14:6c:81:aa:63, IRQ 10
---------- Post added at 03:37 PM ---------- Previous post was at 11:17 AM ----------
I posted too soon. I cannot prevent the network loss by running the Router Statistics monitor after all.
---------- Post added at 07:28 PM ---------- Previous post was at 03:37 PM ----------
Well I now have one instance of the problem under F17, so maybe it is a flaky router???