Yesterday i got an outage on my firewall (Netgate 3100), it did just stop working, internet down and i could not reach the firewall. a fysical restart did solv the problem but this should of course not happen.
in system logs i can se only one thing before restart and that is 5 lines of:
[zone: mbuf_cluster] kern.ipc.nmbclusters limit reached
Here is a screenshot of the log from that time. anyone have a ideah what it means and if it can be related to the problem?
i have tried google but lot of old posts, nothing easy to understand or in german haha.
Not something I have encountered but there is more information here on adjusting that parameter.
thanks, and yes it does looks like the Mbufer did max out and that it is going to max out again later today. The curve follow the exact same pattern as yesterday. have to try to find out what is cousing this becouse i do not think adding more mbufs going to help in this case.
We did find the problem, an external IT company we work with had set up an Chron job with a keep alive ping out to a azure server and that did add more and more processes for each ping. and that did eat up the buffer. we did remove this and we have a flat and stable usage.
thanks for the awesome information.