I am currently experiencing the same problem as a previous user who's problem was deemed answered by them no longer polling snmp. I would like to get to the bottom of the issue and get some help with my scenario. Our network consists of 700+ Adtran 1224 and 1335 (POE and non POE) We recently installed SolarWinds as our Network Monitoring application. After reviewing previous thread we have disabled Adtran polling via Solarwinds. We had approx. 18 switches lock up over a period of 4 days before disabling in Solarwinds We thought this had corrected the issue then have had 4-8 more lock up over the past 3 days. Wondering if the snmp walk or snmp like DOS attack left its impression on the Adtran in tying up memory resources or CPU. And this just took a few days to set it over the edge and cause it to lockup. We are running the blow FW versions on the majority of these switches and see some bugs in the software release notes where an snmp walk can cause a system lockup. I'm wondering how If this theory makes sense to anyone and what cpu or memory would be a good place to look for compare free memory heap cpu etc to give me a warning that a switch is about to lockup. Seems like a memory error moreso then a cpu but I cannot get in to 1 after it locks up so cannot be sure, Any assistance is appreciated.
"NV1335A-17-08-02-00-E.biz" |
"NONVOL:/NV1335A-17-01-01-00-E.biz" |
a random site switch stats - Would statistic tells me the most? I would think the Total Free at the bottom
#show memory heap
Memory Heap:
HeapFree: 78584816
HeapSize: 96136176
Block Managers:
Mgr Size Used Free Max-Used Overhead
(bytes) (blocks) (blocks) (blocks) (bytes)
0 0 1 3 4 128
1 32 23291 1789 25080 802560
2 96 7834 966 8800 281600
3 224 1438 11 1449 46368
4 480 1383 933 2316 74112
5 992 215 2 217 6944
6 2016 103 0 103 3296
7 4064 94 6 100 3200
8 8160 42 3 45 1440
9 16352 33 5 38 1216
10 32736 11 0 11 352
11 65504 12 0 12 384
12 131040 2 0 2 64
13 262112 2 0 2 64
14 524256 0 0 0 0
15 1048544 0 0 0 0
16 2097120 0 0 0 0
17 4194272 0 0 0 0
18 8388576 0 0 0 0
19 16777184 0 0 0 0
Total Overhead (bytes): 1221728
Total Used (bytes): 6101056
Total Free (bytes): 732896
#show processes cpu
System load: 1sec:7.38% 1min:7.43% 5min:7.40% Min: 0.00% Max: 100.00%
Context switch load: 0.17%
Invoked Exec Time Runtime Load %%
Task Id Task Name PRI STA (count) (usec) (usec) (1sec)
1 Idle 0 W 70846543 2003 924498 92.45
2 Thread Pool 2 W 793 289 0 0.00
3 PC Config 5 S 12847152 1022 49432 4.94
4 PacketRouting 36 W 5212274 15 3010 0.30
5 Timer-00 8 W 49871152 5 1189 0.12
6 Nm01 3 W 0 186296 0 0.00
7 Clock 7 W 300829 29 49 0.00
8 FrontPanel 35 W 1982627 422 8559 0.86
9 con0 37 W 483 11 0 0.00
10 CF Manager 7 W 189961 24 31 0.00
11 PCI Bridge 25 W 988712 4 67 0.01
12 Switch 37 W 4544858 13 646 0.06
13 Stacking 7 W 101948 19 19 0.00
14 SwitchQ 13 W 9356001 16 1021 0.10
15 RSTP 35 W 1224035 13 2714 0.27
16 RouteTableTick 4 W 165421 89 124 0.01
17 OSPF 4 W 233047 32 544 0.05
18 IGMPTick 4 W 100844 35 35 0.00
19 IGMP-Receiver 4 W 0 2485173 0 0.00
turning off snmp polling allowed our network to operate without adtran lockups. Seems older Firmware on the netvanta 1224 and 1335 has a known snmp memory leak issue. Upgrading code on all sites may be the long term solution but for now disabling polling cleared up the issue.
turning off snmp polling allowed our network to operate without adtran lockups. Seems older Firmware on the netvanta 1224 and 1335 has a known snmp memory leak issue. Upgrading code on all sites may be the long term solution but for now disabling polling cleared up the issue.