The system became unresponsive

yon 0

@stephenw10 said in The system became unresponsive:

Hmm, well the only thing that jumps out there is the system loading which seems high. What's using CPU cycles there?

Because pfSense version 25.07.1 kept the CPU usage at 100%, making the system unusable, I upgraded to version 25.11, which at least made things a little better.

yon 0

I have now set kern.ipc.maxsockbuf=4831838208 for the system to run. The default value for FRR is too small.

stephenw10

Hmm, is there a specific bug in FRR you're referring to?

yon 0

@stephenw10 said in The system became unresponsive:

Hmm, is there a specific bug in FRR you're referring to?

I believe it is related to FRR. PFsense's design did not match the actual use of FRR.

It is common for some IPv6 sessions to fail to establish a normal connection or to disconnect after running for a period of time, requiring a restart of FRR to restore functionality.

yon 0

yon 0

IPv6 Unicast Summary (VRF default):
BGP router identifier 10.50.1.254, local AS number 65105 vrf-id 0
Read-only mode update-delay limit: 30 seconds
  First neighbor established: 2025/11/13 12:13:58.682
          Best-paths resumed: 2025/11/13 12:14:28.791
        zebra update resumed: 2025/11/13 12:14:31.610
        peers update resumed: 2025/11/13 12:14:43.940
BGP table version 848715
RIB entries 443446, using 41 MiB of memory
Peers 11, using 144 KiB of memory

Neighbor             V         AS   MsgRcvd   MsgSent   TblVer  InQ OutQ  Up/Down State/PfxRcd   PfxSnt Desc
2602:fed:70:a::2  4      39753     64477        25   848715    0    0 00:21:06       226782        7 
2602:fed:70:ca::  4      39753     64964        27   848715    0    0 00:22:56       226782        7 
2602:fed:70:b::10 4      39753    137154        17   848715    0    0 00:23:02       430595        7 
2602:fed:70:d::2  4      39753    137160        27   848715    0    0 00:23:02       430601        7 
2602:fed:70:a::10 4      39753     62506        26   848715    0    0 00:22:55       464179        7 
2602:fed:ca1:a::2   4      39753     56475        27   848715    0    0 00:23:00       221110        7 
2602:fed:ca1:b::10  4      39753     56475        26   848715    0    0 00:22:52       221110        7 
2a0d:24:512:a::2   4      39753     62506        26   848715    0    0 00:22:55       464179        7          fr bgp VP
2a0d:24:513:a::20  4      39753   1122273        27   848715    0    0 00:23:02      2361741        7       FRAVP
2a0d:24:513:b::2   4      39753    726320        27   848715    0    0 00:23:02      2361741        7 
2a0d:24:513:d::2   4      39753    726326        27   848715    0    0 00:23:02      2361741        7         FRA1 WG

Total number of neighbors 11

yon 0

stephenw10

Hmm 4.2M mbufs seems extremely high.

yon 0

@stephenw10 So I think there's a problem.

stephenw10

Well it far higher than even the 1M default we usually set and that is generally far bigger than it needs to be. But you also show only 1400 states which is nothing.

If you exhaust the mbufs that would definitely cause a problem. But you should also see that logging an error.