Description

PTX10001-36mr Junos: 22.2R3-S2.5-EVO reporting the following messages after a link flap.

 

kernel: %KERN-4: [ttp_encap_xmit:1207] JTD Cache skb failed

 

 

Symptoms

After link flap, JTD Cache skb failed messages are seen.


mib2d[15824]: %DAEMON-4-SNMP_TRAP_LINK_DOWN: ifIndex 528, ifAdminStatus up(1), ifOperStatus down(2), ifName et-0/2/4

 mib2d[15824]: %DAEMON-4-SNMP_TRAP_LINK_DOWN: ifIndex 531, ifAdminStatus up(1), ifOperStatus down(2), ifName et-0/2/5

 mib2d[15824]: %DAEMON-4-SNMP_TRAP_LINK_DOWN: ifIndex 604, ifAdminStatus up(1), ifOperStatus down(2), ifName ae12

 rpd[17115]: %DAEMON-4-RPD_RSVP_NBRDOWN: RSVP neighbor x.x.x.x down on interface ae12.0 nbr-type Direct, triggered by IGP neighbor down event

 ifmand[14826]: %USER-3-IFMAND_SPEED_BW_ERROR: UpdateLinkState: bundle ae12: bundle IFD minimum bandwidth or minimum links not met,Bandwidth (Current : Required) 0 : 1 Number of links (Current : Required) 0 : 1

 kernel: %KERN-4: [ttp_encap_xmit:1207] JTD Cache skb failed

 kernel: %KERN-4: [ttp_encap_xmit:1207] JTD Cache skb failed




 rpd[17115]: %DAEMON-4-RPD_RSVP_BACKUP_DOWN: Backup for protecting LSP Jnrp down, using bypass Bypass->x.x.x.x->x.x.x.x, reason: Deleting protected LSP

 mib2d[15824]: %DAEMON-4-SNMP_TRAP_LINK_DOWN: ifIndex 528, ifAdminStatus up(1), ifOperStatus down(2), ifName et-0/2/4

 mib2d[15824]: %DAEMON-4-SNMP_TRAP_LINK_DOWN: ifIndex 531, ifAdminStatus up(1), ifOperStatus down(2), ifName et-0/2/5

 mib2d[15824]: %DAEMON-4-SNMP_TRAP_LINK_DOWN: ifIndex 604, ifAdminStatus up(1), ifOperStatus down(2), ifName ae12

 rpd[17115]: %DAEMON-4-RPD_RSVP_NBRDOWN: RSVP neighbor x.x.x.x down on interface ae12.0 nbr-type Direct, triggered by IGP neighbor down event

 ifmand[14826]: %USER-3-IFMAND_SPEED_BW_ERROR: UpdateLinkState: bundle ae12: bundle IFD minimum bandwidth or minimum links not met,Bandwidth (Current : Required) 0 : 1 Number of links (Current : Required) 0 : 1

 kernel: %KERN-4: [ttp_encap_xmit:1207] JTD Cache skb failed

Solution

The SWERR in question will arise when an application starts sending data on a newly established TCP connection before JTD has been provisioned with the required TCP flow information by FIBD (FIBDv1)/nsagentd (FIBDv2)+ADs- this is not a bug but a perfectly feasible timing-related situation which will
simply result in the outgoing packet being dropped (TCP will subsequently retransmit it). So there is no actual bug here other than the SWERR.

From 22.3-EVO onwards, a series of changes were made to phase out logs and move to error/non-error based statistics. 

 

Modification History

2024-08-21 : Article Created