Router processor loads on 100% because of management process

RouterOS General
1

Hi!
I have an rb450g with license 5, and when I verify the tool/profile I discover that the management process is taking 70% of processor can someone tell me what that mean and how to solve this issue.
I found that this process is from router internal processes but which processes I don’t know.
I’m not running so much on this router about 15 firewall rules and about 15 nat rules, also I use a web proxy.

If anyone know how to solve this please let me know.

2

I found that SNMP access from a running “Dude” can cause very high management CPU load, this was also discussed in the forums after some of the 5.x series patches.

3

Is not because of SNMP even if I turn off the dude the management process is still about 50-80%

4

Hi deejaylight:
what’s Router OS version are you using?
I have the same behavior using 5.12 but not using 5.2 or 4.17 with similar configurations in x86 plataforms monitored with dude 4beta2. You need to disable snmp/routeros access from dude, after that, management process consumed aronud 4-5%.

5

I have this problem too, sent supout.rif to mikrotik

5.7 I didn’t see this.

Upgraded to 5.12 and within a couple days or less CPU gets stuck at 100%

you say disabling SNMP/routeros access from dude?
What do you mean?

In Dude for the device Im monitoring, disable the routeros box? and that router CPU will go back to normal?

also x86 but not running the beta Dude. Just 3.6

EDIT: you were right, as soon as I clicked off the routeros box in Dude, CPU went to normal.

6

i’m using ROS 5.13 x86 (but same happened with 5.11 & 5.12, not in 4.x)

if i enable snmp in ros, cpu goes to 100%… i’m using the dude to monitor this router, but it doesn’t matter if RouterOS box is checked or not in thedude, only matters if snmp is enabled or not in the x86.

7

The same happens here, ROS 5.2
Is it maybe fixed in 5.14?

8

Same Problem on RB1100AHx2.
Management ~50%, total CPU utilization ~70%-90%
Very slow response in Winbox (one minute until ip->address shows addresslist), websites slow, sometimes timeout etc.
After i deleted the device from Dude, from one second to another, everything was fine. Management ~0.5%, very fast response (as expected in fiber backbone) no problems with websites etc.

Will there be a solution some time? There are posts from 2011 with this problem but i could not recognize any reaction from MikroTik

Regards
Manuel Ritter

9

I turned off SMNP on all routers, I never needed it, and the problem has never come back.

10

That is because this is a community support forum, not an official channel for support. The best thing you can do in that kind of situation is generate a supout.rif when the CPU is loaded like that and send it to MikroTik support.

11

At times generating supout.rif with the cpu at 100% is not possible. I tried it so that I could send it to support but the router kept rebooting. Changing it was way easier.

Sent from my GT-P7510 using Tapatalk

12

Its wise having a team from Mitrotik look into this community forum just to give them
an Idea how people feel about their products and services. Truth be told there is a
problem with either the latest boards or firmware and its costing some of us money
due to unnecessary downtimes. On another note generating supout.rif on a 100%
processor is not easy.

Sent from my GT-P7510 using Tapatalk

13

For those interested, this is the response for MT support:

Thank you very much for the clarification.
Yes it is correct, The Dude might cause higher loads (it is known problem and fix
will be included in the next Dude version), you can make workaround and increase
“Mac Mapping Refresh Interval” at Settings → Misc, set it to highest value.