Loading...
Loading...
MTM process crash when large number of MAC addresses in the network
There are a large number of Kubernetes services in the customer network. Pods are frequently created and deleted in the network. Each container uses a different source MAC address. N3K can learn 20,000+ different MAC every day. The memory occupied by the MTM process will continue to increase until the MTM process crashes: N3K-C3548P-10GX(config-if)# show processes mem | in PID|mtm PID MemAlloc MemLimit MemUsed StackBase/Ptr Process 19212 620085248 0 917118976 ffc6b400/ffc6aa40 mtm %SYSMGR-SLOT1-2-SERVICE_CRASHED: Service "mtm" (PID 19212) hasn't caught signal 6 (core will be saved). There have no core file. N3K-C3548P-10GX and N3K-C3548P-XL both have this issue.
N3K learn large number of different MAC every day(old MAC address aged out).
NA
"MTM_PD_MEM_mtm_pd_mac_loc_t " and "MTM_PD_MEM_mtm_pd_mac_entry_node_t " two modules take up a lot of memory space: N3K-C3548P-10GX(config-if)# slot 1 show system internal mtm mem-stats detail Private Mem stats for UUID : Mac Table Manager PD(622) Max types: 30 -------------------------------------------------------------------------------- TYPE NAME ALLOCS BYTES CURR MAX CURR MAX 23 MTM_PD_MEM_mtm_pd_mac_loc_t 1538826 1538826 184659120 184659120 24 MTM_PD_MEM_mtm_pd_mac_entry_node_t 3077652 3077652 123106080 123106080
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.