dr4g0n Posted November 3, 2016 Share Posted November 3, 2016 Hello, unRAID peoples. I just switched my motherboard out for one with a couple more PCIe slots, for some VM fun, but after reinstalling everything and booting up, my add-in, two-port NIC wont grab an DHCP address, it just pulls an APIPA. I (foolishly) deleted the network.cfg, rebooted, and tried to configure my dual-nic as the bond (eth0 and eth1), wich does the APIPA thing, and the onboard NIC (eth2) as a "management port" (DHCP reserved IP assians correctly) ...which worked.. and works through reboots. Heres the current ifconfig: Pastebin cause character limit; http://pastebin.com/uPeHb6wH and heres the network.cfg (I could Not find anything online/here that tells me how to use this to fix the problem..) root@Horde:~# cat /boot/config/network.cfg # Generated settings: IFNAME[0]="br0" BONDNAME[0]="bond0" BONDNICS[0]="eth0 eth1" BONDING_MODE[0]="1" BONDING_MIIMON[0]="100" DHCP_KEEPRESOLV="no" BRNAME[0]="br0" BRNICS[0]="bond0" BRSTP[0]="no" BRFD[0]="0" DESCRIPTION[0]="" USE_DHCP[0]="yes" NETMASK[0]="255.255.0.0" GATEWAY="10.10.0.1" MTU[0]="" IFNAME[1]="eth2" DESCRIPTION[1]="" USE_DHCP[1]="yes" MTU[1]="" SYSNICS="2" My network is thus: Router: 10.10.0.1 Netmask: 255.255.0.0 DNS: 10.10.1.222, 10.10.0.1 Desired IP for the bond: 10.10.1.1 Current IP for the onboard NIC: 10.10.1.2 Servers are in the (sudoish) subnet of 10.10.1.0 Infrastructure (Router, APs) are 10.10.0.0 Clients (my PC, phone, etc; all DHPC reserved) are 10.10.2.0 DHPC guests are 10.10.3.0 Ive got the diag zip if someone needs it, let me know. Hope I can get some help. PS: after the upgrade, everything else works; array, docker, VMs.. Link to comment
John_M Posted November 3, 2016 Share Posted November 3, 2016 Post your diagnostics zip and your DHCP server configuration. Link to comment
Msan Posted November 3, 2016 Share Posted November 3, 2016 Netmask 255.255.0.0 ? should that not be 255.255.255.0 Link to comment
John_M Posted November 3, 2016 Share Posted November 3, 2016 Netmask 255.255.0.0 ? should that not be 255.255.255.0 A 16-bit netmask is a valid configuration, if that is indeed the OP's intention. It's difficult to guess. Link to comment
Msan Posted November 3, 2016 Share Posted November 3, 2016 Netmask 255.255.0.0 ? should that not be 255.255.255.0 A 16-bit netmask is a valid configuration, if that is indeed the OP's intention. It's difficult to guess. Yes it's valid.. not really used in a home environment though usually.. Link to comment
John_M Posted November 3, 2016 Share Posted November 3, 2016 It has to be his intention IMO, otherwise the two name servers would be on different subnets and phones and DHCP guests wouldn't see the router. We need to wait for the OP to respond but I think he's just using IP addresses "wastefully" for his own convenience. There's nothing wrong with that. EDIT: That said, some consumer equipment doesn't allow the setting of less than 24-bit netmasks. I have an old ADSL modem/router that doesn't, for example. Link to comment
dr4g0n Posted November 3, 2016 Author Share Posted November 3, 2016 Netmask 255.255.0.0 ? should that not be 255.255.255.0 Yeah, it's on purpose. I have quasi-subnets, I guess. just using the 3rd octate to designate what the clients are, ie. server, user, guest, etc. Post your diagnostics zip and your DHCP server configuration. Sure. Zip is attached, heres the DHCP: http://puu.sh/s5nSu.png EDIT: That said, some consumer equipment doesn't allow the setting of less than 24-bit netmasks. I have an old ADSL modem/router that doesn't, for example. Well, the card worked fine with the other motherboard, same config. and all the other devices I have currently have no problem with it. horde-diagnostics-20161102-2036.zip Link to comment
dr4g0n Posted November 3, 2016 Author Share Posted November 3, 2016 It has to be his intention IMO, otherwise the two name servers would be on different subnets and phones and DHCP guests wouldn't see the router. We need to wait for the OP to respond but I think he's just using IP addresses "wastefully" for his own convenience. There's nothing wrong with that. I have a Rasberry Pi running Pi-Hole as an ad blocking DNS server, that's what 10.10.1.222 is. It forwards valid DNS to OpenDNS's servers. I have the router (10.10.0.1) as the other just in case the Pi doesn't pick up a local device .local name. Link to comment
John_M Posted November 3, 2016 Share Posted November 3, 2016 The DHCP server whose configuration page you linked allocates addresses from the range 10.10.3.100/16 to 10.10.3.254/16 so how do you expect your unRAID server to receive the specific address 10.10.1.1? Is there another configuration page that reserves IP addresses for MAC addresses, because if there is it's outside the range specified for the DHCP server address pool. Try removing the bond and the bridge and seeing if eth0 and eth1 then work as expected. If they do then add the bridge back and test again, then recreate the bond and test again. If all that fails try using a static address. Link to comment
dr4g0n Posted November 4, 2016 Author Share Posted November 4, 2016 The DHCP server whose configuration page you linked allocates addresses from the range 10.10.3.100/16 to 10.10.3.254/16 so how do you expect your unRAID server to receive the specific address 10.10.1.1? Oh yeah, thats on a different page. But even with no reserved address, the bond wont pick up an ip from the 10.10.3.100-254 range. Ill go ahead and try removing the bridge and bond, see if that works. Link to comment
dr4g0n Posted November 8, 2016 Author Share Posted November 8, 2016 Alright so. I just removed the add in card. I assume my problem is because both the mobo and card are realtek chips. I have an Intel card in the mail. But, now with only one interface.. I can pick up the DHCP address fine, and the local network works great. I can access samba and the webgui fine. But. Im getting 100% dropped packets when trying to connect outside my LAN. So.. yeah, not fun. Link to comment
John_M Posted November 8, 2016 Share Posted November 8, 2016 Im getting 100% dropped packets when trying to connect outside my LAN. So.. yeah, not fun. That has to be a routing problem. Link to comment
dr4g0n Posted November 10, 2016 Author Share Posted November 10, 2016 That has to be a routing problem. Guess so. Bypassed my router and plugged directly into my modem. connects fine.. Link to comment
ken-ji Posted November 10, 2016 Share Posted November 10, 2016 Most DHCP servers and clients are not configured to assign two or more subnets to a single host without colliding on the default route. (And I just noticed the diagnostics didn't include networking stuff like routing table) While the server is in the normal network location/connection run the following commands and post them back ip addr ip route I'm guessing your default route is wrong or clobbered. Link to comment
dr4g0n Posted November 10, 2016 Author Share Posted November 10, 2016 root@Horde:~# ip addr 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/32 scope host lo valid_lft forever preferred_lft forever inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever 2: tunl0@NONE: <NOARP> mtu 1480 qdisc noop state DOWN group default qlen 1 link/ipip 0.0.0.0 brd 0.0.0.0 3: gre0@NONE: <NOARP> mtu 1476 qdisc noop state DOWN group default qlen 1 link/gre 0.0.0.0 brd 0.0.0.0 4: gretap0@NONE: <BROADCAST,MULTICAST> mtu 1462 qdisc noop state DOWN group default qlen 1000 link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff 5: ip_vti0@NONE: <NOARP> mtu 1364 qdisc noop state DOWN group default qlen 1 link/ipip 0.0.0.0 brd 0.0.0.0 6: eth0: <BROADCAST,MULTICAST,PROMISC,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master br0 state UP group default qlen 1000 link/ether d0:50:99:ac:5e:6e brd ff:ff:ff:ff:ff:ff 58: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000 link/ether d0:50:99:ac:5e:6e brd ff:ff:ff:ff:ff:ff inet 10.10.1.3/16 brd 10.10.255.255 scope global br0 valid_lft forever preferred_lft forever root@Horde:~# ip route default via 10.10.0.1 dev br0 metric 258 10.10.0.0/16 dev br0 proto kernel scope link src 10.10.1.3 metric 258 127.0.0.0/8 dev lo scope link root@Horde:~# Well the thing is that this is how Ive had my network set up for over a year, and its worked for every device so far. I also had the server running fine with three NICs and two IPs, one through the br0 with two NICs and the other as sort of a management port via the mobo NIC. it was only after I switched motherboards that I started having this problem. Im using a NETGEAR ProSafe 8 Port VPN Firewall FVS318Gv2 as a router. Link to comment
ken-ji Posted November 10, 2016 Share Posted November 10, 2016 In your current config I don't see anything that could be broken. oh well. Link to comment
John_M Posted November 10, 2016 Share Posted November 10, 2016 In your current config I don't see anything that could be broken. oh well. Current config only shows eth0. Link to comment
dr4g0n Posted November 10, 2016 Author Share Posted November 10, 2016 That's all I have right now. I pulled the two other Nics out to figure out what was up. Link to comment
kode54 Posted November 11, 2016 Share Posted November 11, 2016 Maybe your router is not accepting broadcasts to the correct IP address? It seems a little odd that you've set up a /16 subnet, and while I know even /8 is acceptable for 10.*, it may be over provisioning, unless you really do have 65535 devices in your network. Have you checked any other Linux machines on the network to see if they're broadcasting out properly? Link to comment
dr4g0n Posted November 11, 2016 Author Share Posted November 11, 2016 Maybe your router is not accepting broadcasts to the correct IP address? It seems a little odd that you've set up a /16 subnet, and while I know even /8 is acceptable for 10.*, it may be over provisioning, unless you really do have 65535 devices in your network. Have you checked any other Linux machines on the network to see if they're broadcasting out properly? I only have it /16 because I'm too lazy to make it a /24 and make Actual subnets, so yeah. But when I'm home ill check out broadcast stuff. Link to comment
dr4g0n Posted November 14, 2016 Author Share Posted November 14, 2016 GUYS I SOLVED IT. It was the friggen MTU. Some guy at my work suggested checking that. Sure enough. The default of 1500.. Is too big for my PPPoE stupid Century Link internet. 1492. Yup PPPoE Will Not accept anything over that. EVERYTHINGS WORKING Thanks for trying guys And here, have some cookies! *hands out cookies* Link to comment
kode54 Posted November 14, 2016 Share Posted November 14, 2016 MTU is usually lower for PPPoE framing. Strange that unRAID did not pick that up in the DHCP advertisement. Link to comment
dr4g0n Posted November 14, 2016 Author Share Posted November 14, 2016 I really didnt look into the DHCP, I just set it manually. Meh, it works now. If i run into a problem with another device I'll check the router. Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.