cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1045
Views
0
Helpful
3
Replies

High CPU utilization and packet drop on GRE tunnel

mahesh18
Level 6
Level 6

Hi Everyone,

We have GRE tunnel between 2 sites.

Users complained about slowness and i checked CPU utilization was too high.

It increased from average of 40 to 70% in last few hours.

Here is tunnel interface setting

MTU 17916 bytes, BW 100 Kbit, DLY 50000 usec,

     reliability 255/255, txload 235/255, rxload 241/255

Input queue: 0/75/4339/0 (size/max/drops/flushes); Total output drops: 89

Other end has 39 drops

Is it ok to have drops when there is large amount of traffic passing the tunnel?.

Need to know what else i should look for?

Are these drops 89 ok to have?

Thanks

MAhesh

1 Accepted Solution

Accepted Solutions

Jouni Forss
VIP Alumni
VIP Alumni

Hi Mahesh,

Though this again is not a subject that would be that close to me but I'll have a try.

What device model are you using at the site where you observer the high CPU utilization? Maybe the device cant handle the amount of traffic using this method?

Can you post the whole output of the command "show interface Tunnel x"?

What kind of WAN connection does this GRE Tunnel use?

Is the bandwith set to anything out of the default 8000kbps both in and out?

I think you can set it with the commands

"tunnel bandwith transmit "

"tunnel bandwith receive "

There also a command

"bandwith "

Whatever you are using this connection for, has there been any changes in the network use between the sites that would explain the increased CPU utilization?

It would seem to me according to the above output that the Tunnel is simply "pushing out and in" as much traffic as it can. Or as much traffic it can push accoring to the interface configuration.

txload 235/255, rxload 241/255

This coupled with the high CPU utilization might naturally explain the drops. Though naturally the CPU utilization is probably the effect of the high tunnel bandwith utilization.

- Jouni

View solution in original post

3 Replies 3

Jouni Forss
VIP Alumni
VIP Alumni

Hi Mahesh,

Though this again is not a subject that would be that close to me but I'll have a try.

What device model are you using at the site where you observer the high CPU utilization? Maybe the device cant handle the amount of traffic using this method?

Can you post the whole output of the command "show interface Tunnel x"?

What kind of WAN connection does this GRE Tunnel use?

Is the bandwith set to anything out of the default 8000kbps both in and out?

I think you can set it with the commands

"tunnel bandwith transmit "

"tunnel bandwith receive "

There also a command

"bandwith "

Whatever you are using this connection for, has there been any changes in the network use between the sites that would explain the increased CPU utilization?

It would seem to me according to the above output that the Tunnel is simply "pushing out and in" as much traffic as it can. Or as much traffic it can push accoring to the interface configuration.

txload 235/255, rxload 241/255

This coupled with the high CPU utilization might naturally explain the drops. Though naturally the CPU utilization is probably the effect of the high tunnel bandwith utilization.

- Jouni

Hi Jouni,

i will check with our other teams to know what data is travelling between two sites.

Thanks

Mahesh

Hi Jouni,

High CPU was due to the fact that 2 servers at different WAN sites were talking to each other.

Thanks

MAhesh

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: