某局点S6520X-HI 内存高达93%问题

/

/

设备: S6520X-HI

问题:作为adcampus中的leaf,内存高达93%,随后业务受损,现场收集诊断和日志后,紧急重启设备后业务恢复了.现场还有一台,89%的.

============================================================

  ===============display memory=============== 

Memory statistics are measured in KB:

Slot 1:

             Total      Used      Free    Shared   Buffers    Cached   FreeRatio

Mem:       2036132   1941016     95116         0         0    292480        5.8%

-/+ Buffers/Cache:   1648536    387596

Swap:            0         0         0

LowMem:    1520036   1425788     94248        --        --        --        6.2%

HighMem:    516096    515228       868        --        --        --        0.2%

%@3911738%May 10 16:30:02:072 2023 TSG-2F-leaf-172.18.108.51 DIAG/1/MEM_EXCEED_THRESHOLD: Memory early-warning threshold has been exceeded.

Memory statistics are measured in KB:

         Total     Free      FreeRatio

Mem:     2036132   112292    5%

LowMem:  1520036   111592    --      

HighMem: 516096    868       --      

Free-memory thresholds:

Minor: 5%

Severe: 3%

Critical: 2%

Normal: 6%

Early-warning: 10%

Secure: 15%

Process info(KB):

JID      Used     Name

550      663556   portsecd

616      289144   bgpd

305      187280   xmlcfgd

209      67312    ifmgr

3422961  51548    comsh

Slub info(KB):

Used      Name

286718    kmalloc-8388560

102389    IPCIM_ENTRY_IPV6_cachep

65535     kmalloc-67108816

51060     kmalloc-2048

24575     kmalloc-4194256

%@3942083%May 12 11:29:01:745 2023 TSG-2F-leaf-172.18.108.51 DIAG/1/MEM_EXCEED_THRESHOLD: Memory critical threshold has been exceeded.

Memory statistics are measured in KB:

         Total     Free      FreeRatio

Mem:     2036132   49844     2%

LowMem:  1520036   48976     --      

HighMem: 516096    868       --      

Free-memory thresholds:

Minor: 5%

Severe: 3%

Critical: 2%

Normal: 6%

Early-warning: 10%

Secure: 15%

Process info(KB):

JID      Used     Name

550      667968   portsecd

616      282176   bgpd

305      187280   xmlcfgd

209      67312    ifmgr

540      51532    routed

Slub info(KB):

Used      Name

286718    kmalloc-8388560

103780    IPCIM_ENTRY_IPV6_cachep

65535     kmalloc-67108816

51060     kmalloc-2048

28671     kmalloc-4194256

经分析代码确认是平台的一个已知问题，目前有F6638Pxx版本可以解决:vlan下配置nd snooping和ipsg，两个端口打入相同NS模拟用户频繁迁移，一段时间后停止流量，删除nd snooping后部分ipv6 IPSG表项残留

在出现内存比较低时，先尝试用下面的命令恢复下

[H3C-probe]process restart name ipcimd

Manually restarting a process might severely affect device operation. Perform this operation only under the guidance of H3C engineers. Continue? [Y/N]:y

Restarting process ipcimd[300] on slot 4...

Succeeded.

[H3C-probe]

升级F6638Pxx版本或者使用[probe]process restart name ipcimd手工释放规避.