x86 ROS Device SSL process spikes router CPU to 100%

I am running the Dude on a VM that has a 2.5ghz processor assigned to it.

At frequent intervals ABOUT 30-60 seconds apart (but not exact) the SSL process spikes the CPU to 100%. This seems to effect SNMP probes primarily, as we don’t drop pings but we lose probes and therefor generate a false notification.

A certificate log entry is made called ‘start CRL update’


We have no certificates installed on this machine.

I disabled api-ssl and www-ssl under IP>Services.

There’s no hotspot running a web server on the device.


We are runing v6.26

Please contanct support@mikrotik.com regarding this problem. And includ supout file from that problematic device generated during moment when cpu load has spiked. And one from The Dude server host also.

http://wiki.mikrotik.com/wiki/Manual:Support_Output_File

I’m having the same issue on version 6.37.2.
x86 with 2 CPU, 1800mhz.
CPU is always at 100%, ssl process takes all resources.

Does fix exist?

I’m having the same problem, yours solved?

My problem persists, does anyone have the solution?

Should be fixed in 6.38
What version are you using ?

I am using version 6.39. I have already tried with version 6.38 and also did not work, the SSL gets very high and the CPU goes to 100%. I am using a virtual machine and running on x86.

Please write to support@mikrotik.com
We will need a remote access.
Thanks.

6.39rc13 problem persists.

Aruba CHR OVPN server:

/interface ovpn-server server> /system resource monitor
          cpu-used: 100%
  cpu-used-per-cpu: 100%
       free-memory: 885812KiB



/system resource print
                   uptime: 1w3d3h13m44s
                  version: 6.40.5 (stable)
               build-time: Oct/31/2017 13:05:15
              free-memory: 864.9MiB
             total-memory: 972.9MiB
                      cpu: Intel(R)
                cpu-count: 1
            cpu-frequency: 1699MHz
                 cpu-load: 100%
           free-hdd-space: 19.8GiB
          total-hdd-space: 19.9GiB
  write-sect-since-reboot: 523672
         write-sect-total: 523673
        architecture-name: x86_64
               board-name: CHR
                 platform: MikroTik
/interface ovpn-server server> /tool profile
NAME                    CPU        USAGE
ppp                                 0.5%
ethernet                            0.5%
ssh                                 0.5%
networking                          0.5%
logging                             0.5%
ssl                                  97%
unclassified                        0.5%
total                               100%

Disabling of OVPN server not work - 100% CPU.

/ip service print
Flags: X - disabled, I - invalid
 #   NAME                                       PORT ADDRESS                                                                         CERTIFICATE
 0 XI telnet                                    23
 1 XI ftp                                          21
 2 XI www                                          80
 3   ssh                                       22
 4 XI www-ssl                                     443                                                                                 none
 5 XI api                                        8728
 6   winbox                                     8291
 7 XI api-ssl                                    8729                                                                                 none

wtf??

I been trying to find why my networks have a cyclic interference issue that make them operate like a yo-yo (up-down) and we were thinking it was a feedback issue. At least now I can see other people are experiencing something too! This problem actually cuts clients off from the internet momentarily. To resolve the issue I un-plugged the Dude router and am not using a monitor system. I don’t like flying blind and need a monitor system that works. Please fix this problem ASAP.

Running Dude from a RB750gr3 version 6.39.2 on a microSD chip. All monitored network routers are running 6.39.2 as well. I would like to update to a later version that doesn’t have this problem but I don’t know how to get the Dude to update. Is there a list of commands available?

cgood “SLL CPU 100%” is not fixed yet I m also searching for solution.
in 6.38rc51
*) Fixed - Changed SSL connection timeouts. SSL connections caused CPU overload
but I m running ver. 6.40.4 SSL problem still persist.

you just need to create new SSL certificate. because the certificate you have created using CPU 100% .

try this script


/certificate
add name=ca-template common-name=myCa key-usage=key-cert-sign,crl-sign
sign ca-template name=myCa

6.41.2
RB750Gr3 running Dude
While making pptp-out connection attempts router has cpu ssl spikes up to 40%
When connected, CPU shows 1-2 %, no any spikes

Is there any solution for this problem? I have the same issue with 6.43.2 (I know it is a little bit old ROS version, but the last change with the dude was in the 6.43). I am using the dude with a x86 VM (under proxmox), the CPU usage is 0% until I start the dude. After that all CPUs are 100% loaded and the / tool prifle shows that as many ssl processes go 100% as many CPUs I have. The dude client can’t connect if I want to use secure connection. If I select the plain connection, the client can connect, but the server still has 100% CPU usage.