Prime 3.4 Notifications

[craiglebutt]

Hi

Been using Prime for years, but after we've just had a refresh on WLC and carrying out testing, we found we wasn't getting Notifications from Prime as fast as we should, some times took an hour+ to come through.

Just installed extra 5520 on 8.5.140, all configured to send syslogs to Prime and Kiwi Syslog.

WLC are added to Prime.

When carrying out testing, to make sure we receive logs for HA failover, AP disconnected, ect, these alerts where intermittent.

Could see some alerts on the Alarm Summary, but no alerts.  

All critical and major alerts are setup to email me directly.

I see Prime probes WLC every 5 minutes, so this ties in with why 5 minute delay on some alerts.

But if it is a syslog as well, surely we should get email alerts come through faster?

Any help appreciated.

[omz]

Hi 

I suspect an issue with resources. Is it a vm or physical box?

You can use ncs run test iops command to check the resources.

From shell, you can check what whats taking up resources. 

top

df -kh
free -m

vmstst -s

cat /proc/meminfo

Maintain Prime Infrastructure Server Health

Not sure if the above helps if you been using Prime for years. I would suggest a tac case.

[renjithg]

Please check VM resource reserved to 100% to what allocated. Run the command on Prime shell cli and validate.

-bash-4.1# vmware-toolbox-cmd stat cpures
0 MHz
-bash-4.1#
-bash-4.1#
-bash-4.1# vmware-toolbox-cmd stat memres
0 MB

if VM is professional, CPU needed is 16MHZ and RAM 24GB. Following command above must return 32000MHZ for 16 vCPU and 24 GB for RAM status provided clock frequency of VMware server must be 2GHZ. basically, 

number of vCPU allocated * clock frequency =  total CPU reservation.

Refer Quick start guide for PI 3.4

https://www.cisco.com/c/en/us/td/docs/net_mgmt/prime/infrastructure/3-4/quickstart/guide/bk_Cisco_Prime_Infrastructure_3_4_0_Quick_Start_Guide.html

Note :  We recommend you reserve 100% of CPU and memory resources for optimal performance.