0

In my mail server I'm getting high TCP reset send between working hours, as shown in this grafic:

SNMP grafic TCP Errors

This is a capture of the server traffic filter by tcp-rst:

enter image description here

When I run "netstat -i" I get this:

Kernel Interface table
Iface   MTU Met   RX-OK RX-ERR RX-DRP RX-OVR    TX-OK TX-ERR TX-DRP TX-OVR Flg
eth0       1500 0  3169579734      0     30 0      2277279957      0      0      0 BMRU
lo        65536 0   2035030      0      0 0       2035030      0      0      0 LRU

At the same time of this TCP resets, the "Average load" of the server get up to 150-200, it is possible that this high load be related to those TCP RST? Also, how can I investigate the source of this resets in Linux?

Regards!

0

The packed you highlighted is from some client to your server on port 443. As it is sent by the client, you would have to investigate in the client.

On the other hand, you mention a load of 150-200, so it is a safe guess to assume that your server can't deliver the content to the client before the client has a timeout, and therefor the client resets the connection.

You should find out what causes the high load. You can run top to display the processes that consume the most CPU. Likely candidates are Apache itself, a database like MySQL or a scripting language like PHP.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.