Troubleshooting segment routing

Troubleshooting SR-MPLS

An SR-MPLS BE tunnel cannot be established

Symptom

The output from the display mpls lsp command on a node shows that the node does not have an outgoing label or the outgoing label is not SR-MPLS allocated when an SR-MPLS BE tunnel is established. For example, the FEC for the egress node is 5.5.5.5/32. The following output shows that no SR-MPLS outgoing label to 5.5.5.5/32 exists on this node, indicating that no SRLSP destined for 5.5.5.5/32 exists.

<Sysname> display mpls lsp

FEC                         Proto       In/Out Label    Out Inter/NHLFE/LSINDEX

12.1.1.2                    Local       -/-             GE0/0/1

Tunnel1                     Local       -/-             NHLFE2

Tunnel10                    Local       -/-             NHLFE1

1.1.1.1/32                  ISIS        16010/-         -

2.2.2.2/32                  ISIS        16020/3         GE0/0/1

2.2.2.2/32                  ISIS        -/3             GE0/0/1

3.3.3.3/32                  ISIS        16030/16030     GE0/0/1

3.3.3.3/32                  ISIS        -/16030         GE0/0/1

4.4.4.4/32                  ISIS        16040/16040     GE0/0/1

4.4.4.4/32                  ISIS        -/16040         GE0/0/1

1.1.1.1/1/4122              SR-TE       -/16030         GE0/0/1

                                        16040

Common causes

The following are the common causes of this type of issue:

·     Physical link failure.

·     SR-MPLS label publishment failed, because the IGP or BGP peer relationship was not established.
The SR-MPLS configuration is missing or incorrect.

Troubleshooting flowchart

Figure 1 shows the troubleshooting flowchart.

Figure 1 Flowchart for troubleshooting SR-MPLS BE tunnel establishment failure

 

Solution

To resolve the issue:

1.     On each node along the SRLSP, execute the display interface brief command to verify that both the physical link state and the data link layer state of each interface on the SRLSP are up.

2.     Verify that an IGP/BGP peer relationship is established correctly and the IGP/BGP configuration is correct on each node that the SRLSP traverses. The troubleshooting procedure depends on the IGP protocol used:

¡     When the IGP protocol is OSPF:

-     Execute the display ospf command to check for the Opaque capable field in the command output. If this field exists, Opaque LSA advertisement and reception capability is enabled in OSPF.  If this field does not exist, execute the opaque-capability enable command in OSPF view to enable opaque LSA advertisement and reception.

-     Execute the display ospf peer command to check the value for the State field. If the value is Full, the neighboring routers are fully adjacent. If the value is not Full, see "OSPFv3 neighbor unable to enter Full state" in OSPFv3 Troubleshooting Guide to troubleshoot the issue.

-     Execute the display mpls lsp command to check for the OSPF LSP. The SR prefix SID for each node is manually assigned to the loopback address. If no such LSP is available, verify that OSPF is enabled on each node by using the ospf area command in loopback interface view or the network command in OSPF area view.

<Sysname> display mpls lsp

FEC                         Proto       In/Out Label    Out Inter/NHLFE/LSINDEX

1.1.1.9/32                  OSPF        16010/-         -

1.1.1.9/32                  ISIS        16010/-         -

2.2.2.9/32                  OSPF        16020/17020     RAGG1.4

-     If the output from the display mpls lsp command also contains the BGP LSP, SRLSP generation might fail because of prefix SID conflict. In this case, execute the peer route-policy command to filter out routes learned from the BGP peer.

¡     When the IGP protocol is IS-IS:

-     Execute the display isis command to check the value for the Cost style field to identify whether the link cost style is wide, compatible, or wide-compatible. If the link cost style is neither of them, execute the cost-style command to change the link cost style.

-     Execute the display isis peer command to check the value of the State field. If the value is Up, the IS-IS neighbor relationship is normal. If it is not Up, see "IS-IS neighbor establishment failure" in IS-IS Troubleshooting Guide to troubleshoot the issue.

-     Execute the display mpls lsp command to check for the IS-IS LSP. The SR prefix SID for each node is manually assigned to the loopback address. If no such LSP is available, check for the isis enable command in loopback interface view on each node.

<Sysname> display mpls lsp

FEC                         Proto       In/Out Label    Out Inter/NHLFE/LSINDEX

1.1.1.9/32                  OSPF        16010/-         -

1.1.1.9/32                  ISIS        16010/-         -

2.2.2.9/32                  ISIS        16020/17020     RAGG1.4

-     If the output from the display mpls lsp command also contains the BGP LSP, SRLSP generation might fail because of prefix SID conflict. In this case, execute the peer route-policy command to filter out routes learned from the BGP peer.

¡     When the IGP protocol is BGP:

-     Execute the display bgp peer command to check the value for the State field. If the value is Established, the BGP session is normal. If the value is not Established, see "BGP peer establishment failure" in BGP troubleshooting Guide to troubleshoot the issue.

-     Execute the display mpls lsp command to check for the BGP LSP. If no such LSP is available, verify that the peer label-route-capability command is configured to enable BGP to exchange labeled routes with a peer or peer group.

<Sysname> display mpls lsp

FEC                         Proto       In/Out Label    Out Inter/NHLFE/LSINDEX

1.1.1.9/32                  OSPF        16010/-         -

1.1.1.9/32                  ISIS        16010/-         -

2.2.2.9/32                  BGP         16020/17020     RAGG1.4

3.     Check the SR-MPLS configuration on each node that the SRLSP traverses:

a.     Verify that SR-MPLS is enabled in IS-IS view, OSPF view, or BGP view. If it is not enabled, execute the segment-routing mpls command to enable SR-MPLS.

b.     Verify that a prefix SID has been configured in loopback interface view. If it is not configured, execute the ospf prefix-sid command in OSPF view or isis prefix-sid command in IS-IS view to configure a prefix SID.

c.     Execute the display segment-routing label-block command to identify whether the prefix SID configured in loopback interface view is within the SRGB label range. If the prefix SID is not within the SRGB range, edit the configured prefix SID.

4.     If the issue persists, collect the following information and contact Technical Support:

¡     Results of each step.

¡     The configuration file, log messages, and alarm messages.

Related alarm and log messages

Alarm messages

N/A

Log messages

N/A

The state of the SR-MPLS TE Tunnel is down

Symptom

The output from the display mpls te tunnel-interface command on the ingress node shows that the SR-MPLS TE tunnel is down.

<Sysname> display mpls te tunnel-interface

Tunnel Name            : Tunnel 1

Tunnel Signalled Name  : tunnel1

Tunnel State           : Down (Main CRLSP Down. Backup CRLSP Down.)

...

Common causes

The following are the common causes of this type of issue:

·     A physical link failure exists on the SRLSPs used by the SR-MPLS TE tunnel.

·     The BFD session for detecting the SR-MPLS TE tunnel is down.

·     The SR-MPLS configuration is missing or incorrect.

·     Incorrect SR-MPLS TE tunnel configuration.

Troubleshooting flowchart

Figure 2 shows the troubleshooting flowchart.

Figure 2 Flowchart for troubleshooting SR-MPLS TE tunnel down

 

Solution

To resolve the issue:

1.     On each node along the SRLSP, execute the display interface brief command to verify that both the physical link state and the data link layer state of each interface on the SRLSP are up.

2.     Identify whether the SR TE tunnel down is caused by a BFD session down.

a.     Execute the display this command in SR MPLS-TE tunnel interface view to check for the mpls bfd, mpls sbfd, mpls tunnel-bfd, or mpls tunnel-sbfd command. If any command is available, go to b.

b.     Execute the display mpls bfd or display mpls sbfd command to check the BFD or SBFD session state.

c.     If the status is down, execute the undo mpls bfd, undo mpls sbfd, undo mpls tunnel-bfd, or undo mpls tunnel-sbfd command to delete BFD/SBFD related commands.

3.     If the BFD/SBFD session is normal or no BFD/SBFD session exists, check the SR-MPLS configuration.

a.     In IS-IS view or OSPF view, check the following configuration to verify that SR-MPLS is supported:

-     When the IGP protocol is IS-IS, execute the display isis command to check the value for the Cost style field to identify whether the link cost style is wide, compatible, or wide-compatible. If the link cost style is neither of them, execute the cost-style command to change the link cost style.

-     When the IGP protocol is OSPF, execute the display ospf command to check for the Opaque capable field in the command output. If this field exists, Opaque LSA advertisement and reception capability is enabled in OSPF. If this field does not exist, execute the opaque-capability enable command in OSPF view to enable opaque LSA advertisement and reception.

b.     If you use prefix SIDs for IP traffic forwarding over SRLSPs, identify whether a prefix SID has been configured in loopback interface view. If it is not configured, execute the ospf prefix-sid command in OSPF view or the isis prefix-sid command in IS-IS view to configure a prefix SID. If you use adjacency SIDs for IP traffic forwarding over SRLSPs, enable adjacency SID allocation in OSPF view or IS-IS view, or identify whether an adjacency SID has been configured on the interface on the SRLSP forwarding path. If it is not configured, execute the segment-routing adjacency enable command in OSPF view or IS-IS view to enable SR-MPLS adjacency SID allocation. You can also execute the isis adjacency-sid or ospf adjacency-sid command in interface view to assign an adjacency SID to an adjacency.

c.     Execute the display segment-routing label-block command to identify whether the prefix SID configured in loopback interface view is within the SRGB label range, and whether the adjacency SID configured in interface view is within the SRLB label range. If the prefix SID is not within the SRGB range or the adjacency SID is not within the SRLB label range, change the configured adjacency SID.

4.     Check MPLS-TE tunnel configuration and perform the following tasks based on the establishment mode of the MPLS-TE tunnel:

¡     Over a static SRLSP—Execute the display mpls static-sr-mpls command on the ingress node of the SRLSP to verify that the ordered list of labels represented by the Out-Label field matches the labels allocated for the nodes that the static SRLSP traverses. If the label sequence in the outgoing label stack on the ingress node does not match the static labels configured on each node along the SRLSP, execute the static-sr-mpls lsp command to change the label sequence in the outgoing label stack on the ingress node.

¡     Over an explicit-path SRLSP—Execute the display explicit-path command on the ingress node of the SRLSP to verify that the IP addresses or SIDs match the IP addresses of the nodes along the SRLSP path or local SIDs. In addition, make sure the SID type specified by the nexthop command in explicit path view on the ingress node is consistent with the prefix SID or adjacency SID type configured in interface view on each node along the SRLSP. This means if a prefix SID is configured on the interface, the SID specified by the nexthop command must also be a prefix SID. If they are inconsistent, execute the nexthop command to change the IP address or SID.

¡     Over a PCE-calculated SRLSP—Check for the mpls te delegation command in MPLS-TE tunnel interface view and execute the display mpls te pce peer command to identify whether the PCC and PCE have established a PCEP session. Use packet capture to verify that the controller (PCE) has performed path updates and that the path is correct. In the captured packets, make sure the adjacency SID or next-hop address sent by the PCE uses strict mode, and the prefix SID or node address uses loose mode. If the PCC and PCE have not established a PCEP session and the captured packets do not meet the above requirements, check the configuration on the controller.

5.     If the issue persists, collect the following information and contact Technical Support:

¡     Results of each step.

¡     The configuration file, log messages, and alarm messages.

Related alarm and log messages

Alarm messages

N/A

Log messages

TE/5/TE_BACKUP_SWITCH