View previous topic :: View next topic |
Author |
Message |
Akaihiryuu l33t
Joined: 08 May 2003 Posts: 794 Location: Columbus, OH
|
Posted: Sat Nov 17, 2007 3:42 pm Post subject: Severe bridge issues. [SOLVED - bad switch] |
|
|
Everything has been working fine for upwards of 6 months. Currently for wireless I am using an Atheros card with Madwifi. My network setup looks like this:
br0 (eth0, ath0), eth0 goes to switch/rest of LAN
eth1: internet
br0 has my internal IP address (192.168.0.1), I also run DHCP internally
About an hour ago, the wired computers suddenly lost all network access. DHCP fails to renew, and even the Linux boxes are getting some kind of bogus 169.254 IP address (which should only happen with Windows clients that fail to renew). If I set their IP's manually they can talk to the server, but DHCP fails. DHCP is still working normally on the wireless clients. dhcpd is set to serve on br0, so it shouldn't matter if wired/wireless, and in fact has been working for 6 months. From what I can see of the configuration, it should all work, there is no technical reason why it should not (server logs indicate that bridge is forwarding normally and should be working). I see it as a technical impossibility at this point, but there is obviously something wrong. As an absolute last resort, I rebooted the server (which I NEVER EVER do except to change the kernel), and still nothing. I'm completely lost, any assistance would be appreciated.
Last edited by Akaihiryuu on Tue Mar 25, 2008 8:35 pm; edited 1 time in total |
|
Back to top |
|
|
Paczesiowa Guru
Joined: 06 Mar 2006 Posts: 593 Location: Oborniki Śląskie, Poland
|
Posted: Sat Nov 17, 2007 6:09 pm Post subject: |
|
|
do you use authoritative mode in dhcp software?
if it's not that, maybe some package update broke it? I can give you later my versions of software (I have the same setup) |
|
Back to top |
|
|
Akaihiryuu l33t
Joined: 08 May 2003 Posts: 794 Location: Columbus, OH
|
Posted: Sat Nov 17, 2007 6:38 pm Post subject: |
|
|
I do use authoritative mode for DHCP, but DHCP itself isn't the problem. Even if I set the IP address manually, it still doesn't work. I have to start a ping to the server, then do another ping back from the server to the wired machine, then that one will communicate. Internet access works at that point, but I still cannot ping across the bridge (no communication with wireless systems even though they can see the server/internet also), and DHCP still fails. This is something in the kernel bridging itself that has to be wrong, but I haven't changed the kernel for quite some time. |
|
Back to top |
|
|
Akaihiryuu l33t
Joined: 08 May 2003 Posts: 794 Location: Columbus, OH
|
Posted: Sat Nov 17, 2007 8:21 pm Post subject: |
|
|
I've confirmed that it is not a hardware issue. If I take the bridge out of the equation (put the IP on eth0) and run dhcpd on eth0, everything works as it did before. Of course, that does mean losing my wireless network. Previously, I had the IP address on br0, and had eth0 and ath0 set to null, and ran dhcpd on br0, that worked perfectly fine for a year, and suddenly stopped this morning when I had made no changes. |
|
Back to top |
|
|
Paczesiowa Guru
Joined: 06 Mar 2006 Posts: 593 Location: Oborniki Śląskie, Poland
|
Posted: Sun Nov 18, 2007 3:01 am Post subject: |
|
|
what is your Code: | grep CONFIG_BRIDGE_NETFILTER /usr/src/linux/.config | ? |
|
Back to top |
|
|
Akaihiryuu l33t
Joined: 08 May 2003 Posts: 794 Location: Columbus, OH
|
Posted: Sun Nov 18, 2007 6:37 am Post subject: |
|
|
Paczesiowa wrote: | what is your Code: | grep CONFIG_BRIDGE_NETFILTER /usr/src/linux/.config | ? |
It's on. But I added the lines to sysctl.conf to disable bridge filtering. Oh, here's something weird. I edited config files to take the bridge out (moved services back to eth0). This basically took my wireless network down. Later I was doing some experimenting, set things back to *exactly as they were before*, and now it works fine again. No explanation...in fact I don't have an explanation for why it wasn't working, because nothing was changed. Oh well. Would it help if I disabled that? I'd like to aovid a recurrence, though I have no explanation for why it stopped working and then started again with no changes (I literally just restored my backed up config files from earlier). |
|
Back to top |
|
|
Paczesiowa Guru
Joined: 06 Mar 2006 Posts: 593 Location: Oborniki Śląskie, Poland
|
Posted: Sun Nov 18, 2007 6:46 am Post subject: |
|
|
are you sure you have Code: | iptables -A FORWARD -i br0 -o br0 -j ACCEPT | in non-working configs? |
|
Back to top |
|
|
Akaihiryuu l33t
Joined: 08 May 2003 Posts: 794 Location: Columbus, OH
|
Posted: Sun Nov 18, 2007 4:39 pm Post subject: |
|
|
I don't have that...but right now it's working and I still don't have that. I try not to use interfaces in iptables rules as a general rule, so my forwarding rules just refer to 192.168.0.0/24 instead of the interface. |
|
Back to top |
|
|
Akaihiryuu l33t
Joined: 08 May 2003 Posts: 794 Location: Columbus, OH
|
Posted: Sun Nov 18, 2007 4:44 pm Post subject: |
|
|
Code: | triforce akai # iptables -L -v
Chain INPUT (policy DROP 20M packets, 3060M bytes)
pkts bytes target prot opt in out source destination
2191K 328M ACCEPT all -- lo any anywhere anywhere
13M 4937M ACCEPT all -- any any 192.168.0.0/24 anywhere
7938K 3213M ACCEPT all -- any any anywhere anywhere state RELATED,ESTABLISHED
1248K 118M ACCEPT tcp -- any any anywhere anywhere tcp dpt:ssh-alt
1306 54371 ACCEPT tcp -- any any anywhere anywhere tcp dpt:telnet-alt
1458K 86M ACCEPT tcp -- any any anywhere anywhere tcp dpt:http-alt
20519 3327K ACCEPT tcp -- any any anywhere anywhere tcp dpt:https-alt
30950 1603K ACCEPT tcp -- any any anywhere anywhere tcp dpt:http-alt
0 0 ACCEPT tcp -- any any anywhere anywhere tcp dpt:ftp
4 192 ACCEPT tcp -- any any anywhere anywhere tcp dpt:3784
Chain FORWARD (policy DROP 4225 packets, 457K bytes)
pkts bytes target prot opt in out source destination
794M 385G ACCEPT all -- any eth1 192.168.0.0/24 anywhere
819M 485G ACCEPT all -- eth1 any anywhere 192.168.0.0/24
Chain OUTPUT (policy ACCEPT 558M packets, 572G bytes)
pkts bytes target prot opt in out source destination |
Code: | triforce akai # iptables -t nat -L -v
Chain PREROUTING (policy ACCEPT 43M packets, 4488M bytes)
pkts bytes target prot opt in out source destination
27 1368 DNAT tcp -- any any anywhere anywhere tcp dpt:4000 to:192.168.0.7
173 8304 DNAT tcp -- any any anywhere anywhere tcp dpt:6112 to:192.168.0.7
7 637 DNAT udp -- any any anywhere anywhere udp dpt:6112 to:192.168.0.7
Chain POSTROUTING (policy ACCEPT 6176K packets, 647M bytes)
pkts bytes target prot opt in out source destination
9130K 594M MASQUERADE all -- any any 192.168.0.0/24 anywhere
Chain OUTPUT (policy ACCEPT 6239K packets, 658M bytes)
pkts bytes target prot opt in out source destination |
Code: | triforce akai # brctl show
bridge name bridge id STP enabled interfaces
br0 8000.00146cc49923 no eth0
ath0 |
Code: | triforce akai # brctl showmacs br0
port no mac addr is local? ageing timer
1 00:04:61:9a:00:c4 no 0.00
1 00:13:20:9b:85:f1 no 8.19
2 00:14:6c:c4:99:23 yes 0.00
2 00:19:d2:2d:ba:0a no 7.61
1 00:40:63:da:21:5f yes 0.00 |
Right now everything is working, and this is exactly how it was set up when it wasn't working. In that case, DHCP failed, the computers on eth0 (port 1) could not communicate with anything or ping across the bridge, while the ones on ath0 (wireless, port 2) were working fine. It almost seems like packets were just being dropped, but I don't see any reason why they would be. And I have no explanation for why it started working again with no changes made. |
|
Back to top |
|
|
Akaihiryuu l33t
Joined: 08 May 2003 Posts: 794 Location: Columbus, OH
|
Posted: Tue Mar 25, 2008 8:35 pm Post subject: |
|
|
Well, after lots of intermittent working and lots of pulling my hair out (especially since nothing changed between the working/not working states)...I finally found the source of the problem: my switch. Apparently it was dying and not consistently passing traffic. It stopped working entirely no matter what I did...and then I noticed that the light for one of the ports stayed on even if all ethernet cables were unplugged from it. Power cycle and it came back the same way. Replaced the switch and everything is fine again. So it was never a problem with the bridging setup at all, it was a flakey switch the entire time! |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|