Topics in Cumulus Networks® 1970-01-01T00:00:00Z 593 394 29 78 0 76 7633154 vrf exec default bash and ssh idle timers 2017-11-18T14:27:22Z 2017-11-17T17:05:54Z Vikram Anumukonda vikram_anumukonda i have mgmt vrf configured but with ntp source interface outside the mgmt vrf, I sometimes execute "vrf exec default bash" to get out of the mgmt vrf and check the ntp status.<br /><br />I have noticed the ssh idle timers are not working when i get out of the mgmt vrf&nbsp; (This is on cumulus 3.4.3.) problem 3 2 1 comment 7615242 Testing redistribute neighbor in VX 2017-11-15T23:15:40Z 2017-10-19T21:28:14Z Tarik tarik_houmaiza Hi Team,<br />I'm interested in Cumulus and tried a GNS3 setup with VX to test redistribute neighbor.<br />I'm using pure L3 via OSPF unnumbered and same anycast GW across all leafs however I can't even get Host3 to ping Host4 on same Leaf!<br />I can see all the Hosts IPs are getting redistributed as /32s and can reach all Hosts from Spines.<br />Leaf config sample:<br /><blockquote>interface lo<br />&nbsp; address<br />interface swp1<br />&nbsp; address<br />&nbsp; ip ospf network point-to-point<br />interface swp2<br />&nbsp; address<br />&nbsp; ip ospf network point-to-point<br />interface swp6<br />&nbsp; address<br />interface swp7<br />&nbsp; address<br />ip import-table 10<br />router ospf<br />&nbsp; ospf router-id<br />&nbsp; redistribute table 10<br />&nbsp; network area<br />&nbsp; network area<br />route-map REDIST_NEIGHBOR permit 10<br />&nbsp; match interface swp6<br />route-map REDIST_NEIGHBOR permit 20<br />&nbsp; match interface swp7<br />ip protocol table route-map REDIST_NEIGHBOR</blockquote>Am I missing something?<img src="" title="Image https//d2r1vs3d9006apcloudfrontnet/s3_images/1663356/RackMultipart20171019-18107-13k16za-GNS3_inlineJPG1508448476" /> question 7 2 1 comment 7628972 Presicion time protocol over cumulus 2017-11-15T01:18:40Z 2017-11-10T13:42:47Z Edwin Perez edwin_perez_34ys4usjvb3qx Hello, any one know if cumulus linux suport presicion time protocol? question 4 3 1 comment 7628769 Cumulus Linux 3.4.3 is now available! 2017-11-10T01:55:08Z 2017-11-10T01:55:08Z Pete B pete_b_7033546 Read the <a href="" rel="nofollow">release notes</a>&nbsp;to see what we've fixed. If you want to try it out, download Cumulus VX from <a href="" rel="nofollow">our site</a>&nbsp;or from <a href="" rel="nofollow" title="Link:">Vagrant Cloud</a>. update 0 1 1 create 7627578 MLAG Spine Leaf mesh port channels example 2017-11-09T01:19:56Z 2017-11-08T09:06:55Z Alex Kar alex_kar Dear all,<br /><br />I am currently trying to re-create the MLAG Spine-leaf example shown in the net expample cmd:<br /><br />=====<br />cumulus@spine1:~$ net example clag l2-with-server-vlan-trunks<br /><br />Scenario<br />========<br /><br /><br />We want to create a L2-MLAG leaf-spine network with VLAN trunks to the hosts;<br />these networks are traditionally used in virtualization use cases.<br /><br />In this example We create an MLAG peering relationships between spine1/spine2 as<br />well as leaf01/leaf02 on their swp49 and swp50 interfaces.&nbsp; We'll distribute vlans<br />100-500 through the whole network, trunking VLANs 100-120 to host-11 (with vlan<br />100 untagged).<br /><br />You will need to configure spine1/spine2 and leaf01/leaf02, the steps for each<br />pair are very similar.<br /><br />- create the peering; select one switch to be primary and the other secondary<br />&nbsp; - backup-ip is an optional (recommened) IP address that is separately reachable<br /><br />- create VLANs 100-500<br /><br />- create the mlag downlinks on the spines<br />&nbsp; - spine1 and spine2 MUST use the same clag-id for the downlink to the leaf-pair<br /><br />- create the mlag uplinks on the leafs<br />&nbsp; - leaf01 and leaf02 MUST use the same clag-id for the uplink to the spine-pair<br /><br />- configure a host facing interface for clag<br />&nbsp; - leaf01 and leaf02 MUST use the same clag-id for host-11<br /><br />- connect the clag to host-11 to vlans 100-120, with vlan 100 untagged<br /><br />- review and commit<br /><br />==============<br /><br /><img src="" title="Image https//d2r1vs3d9006apcloudfrontnet/s3_images/1670324/RackMultipart20171108-69867-145fg1i-image_inlinepng1510130755" /><br />Using GNS and the sample configs, I am able to managed to set up the 2 mlag pairs<br />1 - Mlag spine pair - mlag id 1<br />2 - Mlag leaf pair - mlag id 100<br /><br />I could see the mlag pairs are set up properly.&nbsp;<br /><br />sample configs<br />spine2# net add clag peer sys-mac 44:38:39:FF:00:00 interface swp3-4 secondary backup-ip<br />spine2# net add vlan 100-500<br />spine2# net add bridge stp treeprio 16384<br /><b>spine2# net add clag port bond bond-leaf-01-02 interface swp1,swp2 clag-id 1</b><br />spine2# net pending<br />spine2# net commit<br /><br /><br />leaf01# net add clag peer sys-mac 44:38:39:FF:00:01 interface swp3-4 primary backup-ip<br />leaf01# net add vlan 100-500<br /><b>leaf01# net add clag port bond bond-to-spines interface swp1-2 clag-id 100</b><br /><b>leaf01# net add clag port bond bond-to-host-11 interface swp5 clag-id 1</b><br />leaf01# net add bond bond-to-host-11 bridge trunk vlans 100-120<br />leaf01# net add bond bond-to-host-11 bridge pvid 100<br />leaf01# net pending<br />leaf01# net commit<br /><br />However, what I am puzzled is, on the leaf switch, there is no need to configure a LACP bond as the dual connected ends of the spine mlag (id 1) pair. Also at each leaf switch, the bond only cover the 2 local uplinks.<br /><br /><img src="" title="Image:" /><br /><br />In theory, shouldnt that be a port channel of the 4 links across the 2 leaf switches (logically)?&nbsp;<br /><br />Vice versa, on the spine(s) ends, shouldnt that be a need LACP bond for the leaf mlag pair (id 100)?<br /><br />Any guidance is much appreciated. Tks question 4 3 1 comment 7627718 Connecting eth0 port causes traffic coming in front panel ports to use eth0 as default gateway. Will enabling vrf mgmt on eth0 solve this? 2017-11-08T14:43:30Z 2017-11-08T14:24:02Z penny01 daniel_pennington We want to setup an OOB management network, but before plugging eth0 in again I have a question regarding enabling vrf management on eth0.<br />I found that when connecting the eth0 port that traffic coming in on the front panel ports will use it as a default gateway. I set the default gateway already to another device and when eth0 gets plugged in the traffic goes through it instead.<br />Reading about vrf management (<a href="" rel="nofollow" title="Link:"></a>) sounds like it could solve this. Is it true to say that enabling vrf mgmt on eth0 would make it use a separate routing table?<br />Would this then keep the front panel traffic from traversing it as a default gateway and use the original default gw configured?<br /><br />Kind regards,<br />penny01 question 2 2 1 comment 7601784 Does Cumulus Linux support SPB (802.1aq)? 2017-11-08T14:05:26Z 2017-09-28T14:39:04Z penny01 daniel_pennington Does Cumulus Linux support shortest path bridging? question 2 2 1 comment 7625716 vnet interfaces on ubuntu 2017-11-06T15:01:12Z 2017-11-05T07:14:13Z shakir shakir_ahmed_ibrahim i am trying to use two interfaces eth1 and eth2 on the same ubuntu server to two different cumulus linux switches (all virtual), how do i enable bridging here on the ubuntu to use vlans 10 and 20.. i had seen this sample config --&nbsp;&nbsp;<code>auto br-10</code><br /><code>iface br-10 inet manual</code><br /><code>&amp;nbsp;&amp;nbsp;</code><code>bridge-ports eth1&#46;10 vnet0</code><br />&nbsp;<br /><code>auto br-20</code><br /><code>iface br-20 inet manual</code><br /><code>&amp;nbsp;&amp;nbsp;</code><code>bridge-ports eth2&#46;20 vnet1&#46;&#46;&#46; would like to know, why do we need vnets here as we are already using two eth interfaces</code> question 1 2 1 comment 7621254 Route Leaking from vrf to global and vice versa 2017-10-31T18:59:56Z 2017-10-29T04:11:06Z Vikram Anumukonda vikram_anumukonda I followed the instructions&nbsp;(&nbsp;<a href="" rel="nofollow" target="_blank" title="Link https//docscumulusnetworkscom/display/DOCS/VirtualRoutingandForwarding-VRF"></a>&nbsp;)&nbsp; to configure&nbsp;route-leaking &nbsp;in cumulus cloud , but couldn't get it to working.&nbsp;<br /><br />I have also added the routes in vrf and global table using the below commands<br />net add routing route vlan2 vrf cust1<br />net add routing route vlan3<br /><br />here is the output from&nbsp;<br /><br /><br />cumulus@leaf01:mgmt-vrf:~$ net show route | grep 100.124.[0-1].0<br />O&nbsp; &nbsp; [110/10] is directly connected, vlan2, 01:28:15<br />C&gt;* is directly connected, vlan2<br />K * is directly connected, unknown inactive<br />S&nbsp; &nbsp; [1/0] is directly connected, unknown inactive<br /><br /><br />cumulus@leaf01:mgmt-vrf:~$ net show route vrf cust1 | grep 100.124.[0-1].0<br />S&gt;* [1/0] via, vlan101<br />K * is directly connected, unknown inactive<br />S&nbsp; &nbsp; [1/0] is directly connected, unknown inactive<br />C&gt;* is directly connected, vlan3<br />C&gt;* is directly connected, vlan101<br /><br /><br /><br />i get a Unreachable when pinging from server01[] to vlan2[] ,&nbsp;<br />cumulus@server01:~$ ping<br />PING ( 56(84) bytes of data.<br />From icmp_seq=1 Destination Host Unreachable<br /><br />is vrf route leaking not supported in Cumulus VX ? question 8 2 1 comment 7619485 OSPF with unnumbered interfaces with multiple areas possible? 2017-10-30T16:27:57Z 2017-10-26T07:16:23Z Rainbow Dash sascha_frey I'm following this article&nbsp;<a href="" rel="nofollow" target="_blank" title="Link https//supportcumulusnetworkscom/hc/en-us/articles/202796476-OSPF-Unnumbered-Sample-Configurations"></a>&nbsp;to setup OSPF with unnumbered interfaces on my&nbsp;CumulusVX test environment.<br />I've successfully connected two routers in area<br />I tried to add another router in a different area, but it doesn't show up as a&nbsp;neighbor.<br /><br />Each router has a /32 on lo and all its interfaces.<br /><br />Is it even possible to use multiple areas with unnumbered interfaces?<br />If yes, what might be wrong?<br /><br /><br />Router "core1" is connected (swp1, swp2)&nbsp;to "core2" (swp1, swp2).<br />Router "aggr1" is connected (swp1, swp2)&nbsp;to "core1" (swp3, swp4)&nbsp;<br /><br />config from core1:<br />&lt;code&gt;<br />interface swp1<br />&nbsp;ip ospf network point-to-point<br />!<br />interface swp2<br />&nbsp;ip ospf network point-to-point<br />!<br />interface swp3<br />&nbsp;ip ospf network point-to-point<br />!<br />interface swp4<br />&nbsp;ip ospf network point-to-point<br />!<br />router-id<br />!<br />router ospf<br />&nbsp;ospf router-id<br />&nbsp;network area<br />&nbsp;network area<br />&lt;/code&gt;<br /><br />Config from aggr1:<br />&lt;code&gt;<br />interface swp1<br />&nbsp;ip ospf network point-to-point<br />!<br />interface swp2<br />&nbsp;ip ospf network point-to-point<br />!<br />router-id<br />!<br />router ospf<br />&nbsp;ospf router-id<br />&nbsp;network area<br />&nbsp;network area<br />&lt;/code&gt; question 5 2 1 comment 7617944 Spine&leaf BGP unnumbered and VRF, subinterfaces on interlinks 2017-10-24T11:59:32Z 2017-10-24T10:45:45Z Jakub Bitenc jakub_bitenc So, I found this "leaf spine vrf bgp" topic: <a href="" rel="nofollow" title="Link:"></a>.<br />I wonder if it's still the only and appropriate way to implement this functionality, which is to let spines and leaves establish BGP neighborship. My fresh experience with Cumulus Linux teaches that something might have changed since the topic mentioned below (it's 7 months old). question 4 3 1 comment 7617232 Error in output 2017-10-24T06:59:18Z 2017-10-23T13:13:44Z Sharon Tubi sharon_tubi line "auto swp55" in /etc/network/interfaces changes to "@@ -255,21 +255,21" after committing changes with NCLU to swp24. Did anyone see it before?<img src="" /> problem 3 3 1 comment 7280546 802.1Q Tunneling (Q-in-Q) Configuration 2017-10-23T21:54:11Z 2016-08-08T17:55:11Z Bryan Martin bryan_martin_2gl8wtno345vs I am using Cumulus Linux 2.5.7. This is a followup question to <a href="" rel="nofollow" target="_blank" title="Link https//communitycumulusnetworkscom/cumulus/topics/ability-to-insert-vlan-id-into-packets">this</a>&nbsp;previous post.&nbsp;<br /><br />I am hoping to insert a VLAN id into packets that are already VLAN tagged&nbsp;<b>without&nbsp;</b>VLAN translation. I would like to preserve the original VLAN id and place an outer VLAN tag into the packet as well.<br /><br />I hope to accomplish this:<br /><br /><img src="" title="Image https//d2r1vs3d9006apcloudfrontnet/s3_images/1463262/RackMultipart20160808-67058-11o8qk7-qinq1_inlineJPG1470677967" /><br /><br />Switch inserts additional VLAN id before original VLAN id:<br /><br /><img src="" title="Image https//d2r1vs3d9006apcloudfrontnet/s3_images/1463264/RackMultipart20160808-93243-mqyrzq-qinq2_inlineJPG1470678061" /><br /><br />Here is what I have tried so far:<br /><br /><b># cat /etc/network/interface<br /></b>auto swp1.100iface swp1.100<br /><br />auto swp2.37<br />iface swp2.37<br /><br />auto brQinQ<br />iface brQinQ<br />&nbsp; &nbsp;bridge-ports swp1.100 swp2.37<br /><br /><b># sysctl -p /etc/sysctl.d/multiple_vlans.conf&nbsp;</b><br />net.bridge.bridge-allow-multiple-vlans = 1<br /><br />With this configuration, traffic passing into swp1 with VLAN 100 egresses out of swp2 with VLAN 37. The original VLAN (100) is lost however. I would like to preserve the original VLAN and place the new VLAN before the original. Is this possible?<br />&nbsp;<br />Additionally, &nbsp;I have tried the following:<br /><br /><b># cat /etc/network/interface<br /></b>auto swp1.100<br />iface swp1.100<br /><br />auto swp2.100<br />iface swp2.100<br /><br />auto swp2.100.37<br />iface swp2.100.37<br /><br /><b>#ifreload -a</b><br />warning: could not determine vlanid<br />error:&nbsp; question 10 11 1 comment 7617101 ONIE SCP, SSH, FTP Enabling 2017-10-23T18:08:31Z 2017-10-23T07:40:56Z Darshan TM darshan_tm I am not able to connect the switch with WinSCP or FTP as I am getting the error as SCP oe FTP is not enabled in ONIE switch. How do I enable those to Install and OS file question 3 2 1 comment 7615478 Add switchname/vlan id to dhcprelay pif-circuit-id 2017-10-23T12:58:14Z 2017-10-20T07:27:30Z Eelco Nieuwstad eelco_nieuwstad I would like to use option 82 to assign devices an ip based on the port on which they are connected. However when I use the pif-circuit-id I only see the switch port a format like <br />switch.port.vlan would make it easier to uniquely identify a device idea 3 2 1 comment