sofo
Hi All,
I am facing a strange problem.
I even do not know how to debug and find out the root cause.
Can you please focus some light , What should I look for. I spent many days without any success.
I have 2 linux openvpn servers (UDP ) ( v 2.3.2).
After running 24-26 hours openvpn servers become non responsive ( in both machine).
and after 2-3 hours they start serving automatically with full speed.
During that down time I can ping the server ip. I can see openvpn server still running ( TOP ) with 0.0 % CPU usage.
I can ping other custom UDP server running on different port in same IP to check if it is UDP packet related problem in the server network.
In LOG there is nothing special (verb 4 )..no shortage of private IPs. I noted CPU usage does not exceed 2-3 % ever.
So Dont think any Attack.
During this time If I try to connect Management server using telnet..It connects but just shows Blank screen only.
Apart from the down time openvpn server are fast and stable and serve unexpectedly.
Pls note: Max simultaneous user is very low ..max 10-15.
But I must say that few days for both server I noticed my log file 50% filled up with following error with same user name. I use user-pass auth.
< packet HMAC authentication failed openvpn >.
normally in a day my log file size goes up to 500 -700 KB.
but when that < HMAC authentication > related error comes I see log file size goes upto 15 MB in a day.
Could it be attack.
I have set
--connect-freq 60 60
--stale-routes-check 60 60
without any luck.
But some time I see server down ( idle ) without this error. I can not relate.
When I use TCPDUM i can see UDP packet coming from client machine but seems openvpn server does not like to accept it.
and as I mentioned after 2-3 hours all just becomes normal.
Restarting machine solves the problem immediately.
Can you pls tell em how to debug this scenario.
Thank you
sofo
Hi all,
After spending 24 hours again in front of my machine ..I can confirm there is not attack or any abnormal behavior.
I can see UDP packets are coming to my box using TCPDUMP...I can ping other custom UDP server binding in same ip different port.
OpenVPN status LOG is OK. Internal routing table is ok. There is nothing abnormal in Log- verb 6.
Just openvpn server not receiving any UDP packet...How can I debug this scenario. Though I can see Openvpn server is running consuming 0% CPU.
Thank you
