High CPU utilization on switchesWe have experienced potential issue when you use SNMP v3 and vFabric support on Brocade switches.
Some switches might have a problem with 60 secs pooling and STOR2RRD SNMP queries cause increased CPU load on switches.
It has been reported on Brocade DCX switches with 7.3.0c, 7.3.1a, 7.4.1.b and 7.4.1d firmware.
Apparently even some Cisco switches might suffer by that like MDS 9509s with 6.2(13a).
Note the issue seems to be very rare, we have got just a few reports.
SolutionIncrease pooling to 5 minutes, just change crontab scheduling to:
0,5,10,15,20,25,30,35,40,45,50,55 * * * * /home/stor2rrd/stor2rrd/load_sanperf.sh >/home/stor2rrd/stor2rrd/logs/load_sanperf.out 2>&1This is standard pooling frequency since version 2.10.
Note on some SAN Directors even 10 minutes samples do not resolve high CPU usage.
It is a known issue community.brocade.com/t5/Fibre-Channel-SAN/DCX-SAN-switches-CPU-utilization-reaching-100/td-p/85500