Hi, I have two active/passive CARPed boxes of 5.9 with latest errata patches which serve as NAT firewalls, but also as pptp servers, which authenticate users from Microsoft's implementation of radius server (NPS).
In last two weeks, I had three occurrences of npppd going down. Last 32650 (!) lines of npppd log are the following two lines: ---cut-here--- Sep 14 15:32:50 nat1 npppd[11097]: sendto() failed in radius_request0: No route to host Sep 14 15:32:50 nat1 npppd[11097]: ppp id=956 radius accounting request failed: No route to host ---cut-here--- Interestingly enough, last two lines have timestamp of only one second later from the first two (of 32650): ---cut-here--- Sep 14 15:32:51 nat1 npppd[11097]: ppp id=956 radius accounting request failed: No route to host Sep 14 15:32:51 nat1 npppd[11097]: sendto() failed in radius_request0: No route to host ---cut-here--- I am redirecting all the npppd logging to separate file by the following configuration in syslog.conf: ---cut-here--- !!npppd *.* /var/log/npppd !* ---cut-here--- So, I would say the system logged 32650 lines of inaccessible radius server in just one second, and finally crashed. I noticed also following message in daemon log from a few minutes before the crash: ---cut-here--- Sep 14 15:29:22 nat1 ospfd[2087]: send_packet: error sending packet on interface bnx0: No buffer space available ---cut-here--- bnx0 is the interface over which radius server is accessed. I am aware of errata 008_bnx.patch, and it is applied on this box as I said earlier. Is there anything I can do to prevent this? Thank you in advance, -- Before enlightenment - chop wood, draw water. After enlightenment - chop wood, draw water. Marko Cupać https://www.mimar.rs/
