CSCed29335 - Higher cpu util on Cat4k Mgmt HiPri and Cat4k Mgmt LoPri after IOS upg

[Kamran Mustafayev]

Hi,

We have 4900 series platform (to be exactly  4 4948 Catalyst Switch)

We upgraded 2 of 4 switch(from SG1 to SG6) before and they worked fine(2 of them in on site geographically separated from 2 other, all of them are in shared pvst ring)

when we upgraded the third switch this problem occured, on two switches upgraded before Cat4k Mgmt HiPri and Cat4k Mgmt LoPri processes utilized more than 50% of CPU, and no traffic flowed through the switch, we cant even ping a neihgbor switch in the same site, we decided to return an oldest version of IOS sg1 and it worked fine as before, cpu utilization backed to normal state

Waiting for your response

Thank you

[Ashish]

Hi Kamran,

Cat4k Mgmt HiPri and LoPri aggregate multiple platform-specific tasks that perform the essential management functions on the switch like control plane and data packets that must be software-switched or processed

Do you have outputs from the time of issue? 'show process cpu sorted', 'show platform health' any debugs you might have taken.

Let us know.

Thanks

[Kamran Mustafayev]

Hi,

I will have them soon

[Kamran Mustafayev]

We put the same type of switch with this sg6 image and it worked fine, it seems to be old switch problem