CRS112-8P-4S w/ ROs 7.12 cpu usages increase by time to time

RouterOS General
1

hi all,

i install couple of brand new crs112-8p-4s came with ros 6.48.6 and upgraded to 7.12.
the configuration quite simple, that are tagged 3 vlan on vlans and egres vlan translation tab. other configuration are powered cycle if ping failure, dhcp-client, ntp-client, romon.
but.. the cpu of all of the switch act the strange behaviour that always reach to 100% usages from time to time. currently i put script to reboot the switch if the cpu reach 90%

here is the graps of the cpu usages from last three days

see the blue one (cpu)

and here is the configuration :

/interface bridge add name=bridge-local
/interface ethernet
set [ find default-name=ether1 ] loop-protect=on poe-voltage=low power-cycle-ping-address=192.168.1.61 power-cycle-ping-enabled=yes power-cycle-ping-timeout=5m
set [ find default-name=ether2 ] loop-protect=on poe-voltage=low power-cycle-ping-address=192.168.1.62 power-cycle-ping-enabled=yes power-cycle-ping-timeout=5m
set [ find default-name=ether3 ] loop-protect=on poe-voltage=low power-cycle-ping-address=192.168.1.63 power-cycle-ping-enabled=yes power-cycle-ping-timeout=5m
set [ find default-name=ether4 ] loop-protect=on poe-voltage=low power-cycle-ping-address=192.168.1.64 power-cycle-ping-enabled=yes power-cycle-ping-timeout=5m
set [ find default-name=ether5 ] loop-protect=on poe-voltage=low power-cycle-ping-address=192.168.1.71 power-cycle-ping-enabled=yes power-cycle-ping-timeout=5m
set [ find default-name=ether6 ] loop-protect=on poe-voltage=low power-cycle-ping-address=192.168.1.72 power-cycle-ping-enabled=yes power-cycle-ping-timeout=5m
set [ find default-name=ether7 ] loop-protect=on poe-voltage=low power-cycle-ping-address=192.168.1.73 power-cycle-ping-enabled=yes power-cycle-ping-timeout=5m
set [ find default-name=ether8 ] loop-protect=on poe-voltage=low
set [ find default-name=sfp9 ] loop-protect=on
set [ find default-name=sfp10 ] loop-protect=on
set [ find default-name=sfp11 ] loop-protect=on
set [ find default-name=sfp12 ] loop-protect=on
/interface bridge port
add bridge=bridge-local interface=ether1
add bridge=bridge-local interface=ether2
add bridge=bridge-local interface=ether3
add bridge=bridge-local interface=ether4
add bridge=bridge-local interface=ether5
add bridge=bridge-local interface=ether6
add bridge=bridge-local interface=ether7
add bridge=bridge-local interface=ether8
add bridge=bridge-local interface=sfp9
/interface ethernet switch egress-vlan-tag
add tagged-ports=ether1,ether2,ether3,ether4,ether5,ether6,ether7,ether8,sfp9 vlan-id=10
add tagged-ports=ether1,ether2,ether3,ether4,ether5,ether6,ether7,ether8,sfp9 vlan-id=22
add tagged-ports=sfp9 vlan-id=50
/interface ethernet switch vlan
add ports=switch1-cpu,ether1,ether2,ether3,ether4,ether5,ether6,ether7,ether8,sfp9 vlan-id=0
add ports=ether1,ether2,ether3,ether4,ether5,ether6,ether7,ether8,sfp9 vlan-id=10
add ports=ether1,ether2,ether3,ether4,ether5,ether6,ether7,ether8,sfp9 vlan-id=22
add ports=sfp9,sfp10 vlan-id=50
/interface ethernet switch set drop-if-invalid-or-src-port-not-member-of-vlan-on-ports=ether1,ether2,ether3,ether4,ether5,ether6,ether7,ether8,sfp9
/ip dhcp-client add interface=bridge-local
/ip dns set allow-remote-requests=yes cache-max-ttl=12h
/snmp community set [ find default=yes ] addresses=192.168.1.1/32 name=Harmoni
/snmp set contact=Felix enabled=yes location="Panel Floor #6" trap-version=2
/system clock set time-zone-name=Asia/Jakarta
/system identity set name="Switch Floor-6"
/system ntp client set enabled=yes
/system ntp client servers add address=id.pool.ntp.org
/tool netwatch add disabled=no down-script="/system reboot" host=192.168.1.1 interval=15m startup-delay=15m test-script="/ping 192.168.1.1" type=simple
/tool romon set enabled=yes

i have another configuration with the same type of switch and with the same configuration, just upgraded to 7.12 and they are not act the same behaviour. it means, if was configured with 6.48.6 and than upgraded to 7.12

can anyone see the problem source ?

thank you
crs112 7-12 cpu usages.png

2

My guess: it’s your “test-script” setting (it’s very probably needless). According to netwatch docs, the netwatch service itself already does ping test by default and test-script property defines additional test to be run after the probe (simple ICMP by default) already finishes.
Since /ping command runs indefinitely by default and netwatch doesn’t stop test-script forcefully, I’m guessing that your CRS starts a new ping every 15 minutes (netwatch interval setting). Since ROS needs some CPU cycles to process that ping command and after a while there will be a number of ping commands running, it will overwhelm the modest MIPSBE CPU in your device.

3

hi all,

down grade to the last long-term; ros 6.49.10 seems solve the problem; no changes on the configuration of switches.
i did downgrade it at 12 noon and now the graph of cpu usage seems to be stable

P
crs112 ros6-49-10.png

4

hi,

the problem start before that mentioned script made. i made that script to prevent the router unaccessible due to cpu reach 100% and could not ping to anywhere (with result time-out).

P

v7.12.1 [stable] is released!
5

i dont see too much benefit upgrading this Old Switches to V7, i hope you dont need to use some V7 feature on that devices

6

well, i do has port security / mac-address filter problem on v6 at the other site and it has been solve when upgraded into 7.12. but… they do not have the same strange memory behaviour. their cpu memory stay low as if they are still on v6

EDIT: the problem seems to be solved after upgrade to 7.12.1. we will see if port security / mac filter address issued exist on this version

P