Killed: out of swap space
-
Hi,
I have 2 Netgate SG-8860 running pfsense version 2.4.3-RELEASE (amd64) configured as a high availability CARP redundancy cluster and one of the 2 firewalls every few days all of a sudden get a out of swap space error and the GUI becomes inaccessible and the entire firewall stops operating and doesn't let traffic through. My system has 8gb of RAM and on average we use 2gb and 0mb of swap space. Attached is screenshots of the RRD data of 2 of the times the error happened. Any suggestions would be much appreciated.
![RRD Data 1.PNG](/public/imported_attachments/1/RRD Data 1.PNG)
![RRD Data 1.PNG_thumb](/public/imported_attachments/1/RRD Data 1.PNG_thumb)
![RRD Data 2.PNG](/public/imported_attachments/1/RRD Data 2.PNG)
![RRD Data 2.PNG_thumb](/public/imported_attachments/1/RRD Data 2.PNG_thumb) -
Need to watch it and look at the processes to see which one is growing.
My guess: it's a package.
You can run ps alxw from Diagnostics > Command Prompt or from the shell. You probably want to find whatever is growing in VSZ and/or RSS.
-
The problem is it only happens when the firewall is the active node and the second it happens it becomes inaccessible and i cannot afford to bring it back online since i am using it for a live production environment.
-
Which packages are you running?
-
You need to watch those graphs and when you see it growing find out what it is before it consumes all of your swap and blows up.
-
I am running haproxy and pfblcokerng.
We monitor the graphs constantly and it never uses more then %25 of the RAM but the second this problem happens we get locked out of the GUI and command line, and it only happens every few days of using it as the MASTER node.
-
pfBlocker has a rep for being a resource hog sometimes. I could disable it for a few days and monitor the status.
-
Ok, thanks for your suggestion will give that a try.