OPERATIONAL DEFECT DATABASE
...
...
MPLS_LSD, IPV4_RIB, or IPV6_RIB process is occupying a large amount of memory (Can be seen with show processes memory) of OSPF is not able to enable the protocol on now added interfaces There is also possibility that during the timing gap of deleting the UG (which fails) there are also rewrite updates, that get stuck in this limbo and never sent down to FIB. As result there LSD/RIB and FIB could get out of sync.
Seen on any XR platform, in releases 5.3.2 and beyond. State can be determined with commands such as described in Further Problem description. Looking for Update Groups (UG) with IDs other than 0,1,2,3 that are in frozen or caught_parent state without being eventually cleaned up.
A process restart of the given process name (mpls_lsd, or ipv4_rib, or ipv6_rib), should remove the memory leak with minimal traffic impact if done during time without major route churn. If there is also sync up problem between tables then restart will clear out such condition as well.
Commands to determine if frozen/stale Update Groups (UGs) are present in LSD or RIB. These commands should be run multiple times with a space of time such as 60seconds between to see they are permanently stuck. LSD: show mpls lsd private v4: num_pulses=1666 UGs Index:0, UG/parent=(1/-2), num_messages=0, num_updates=0, num_caught_parent=0, num_pauses=0, flow ctrl paused:FALSE Index:1, UG/parent=(2/-2), num_messages=0, num_updates=0, num_caught_parent=0, num_pauses=0, flow ctrl paused:FALSE Index:4, UG/parent=(0/-2), num_messages=287911, num_updates=25630616, num_caught_parent=0, num_pauses=141, flow ctrl paused:FALSE Index:5, UG/parent=(7/0), num_messages=1570, num_updates=36864, num_caught_parent=1, <== If these are present continuously and have caught_parent=1 num_pauses=6, flow ctrl paused:FALSE RIB: show rib update-groups Update Group Client ID Parent UG Redist Freeze Count 3 0 0 outsync 0 2 19 N/A insync 0 0 20 N/A insync 0 1 21 N/A insync 0 5 22 0 insync 1 <== If these are present continuously and have freeze count 1 6 23 1 insync 1 <== If these are present continuously and have freeze count 1 alternatively ITAL is showing "V4 UG Merged : FALSE" show ital dbase if-dbase tenGigE 0/7/0/0 process ospf- location 0/RSP0/CPU0 Number of links = 1 Interface Name : TenGigE0_7_0_0 VRF Name : default Handle : 120000c0 VRF ID : 60000000 AF : IPv4 SAF : Unicast Type : Unknown Type BW : 10000000 IPv4 State : 3 IPv6 State : 17 IPv4 MTU : 1500 IPv6 MTU : 0 Last IPv4 event : 0 Last IPv6 event : 0 IPv4 created : no IPv6 created : no LDP state : 1 TE tun type : 0 Null Interface : no FWD ref : yes Mgmt Interface : no V4 UG Merged : FALSE <<<<<<<<<<<<<<<<<<< V6 UG Merged : FALSE Unnum. handle : 0 Pri IPv4 Addr : Addr X.X.X.X, Pre_Len 30, Change_type 1 IPv6 Loc Addr : Addr ::, Pre_Len 0, Change_type 0 For synchronization problem it can be verified if information between lsd/rib and fib tables matches via show lcc ipv4/ipv6 unicast all and sh rcc ipv4/ipv6 unicast all.
Click on a version to see all relevant bugs
Cisco Integration
Learn more about where this data comes from
Bug Scrub Advisor
Streamline upgrades with automated vendor bug scrubs
BugZero Enterprise
Wish you caught this bug sooner? Get proactive today.
