Hello all. Sorry to double-post here after reporting this very same issue to Fedora bugtracker, but I couldn't get any answers there. I'm experiencing crashes on 2 different machines (albeit with similar hardware). These machines run identical diskless (netbooted to custom ramdisk created using livecd-tools) VPN server images. Workload includes accel-ppp (PPTP/L2TP server), bird (OSPF), snmpd (monitoring) and the usual ntp/ssh/etc. All machines run x86-64 on rather old 1- and 2-socketed Intel platforms (S1200BTL or S3210SH), using 4GB RAM and e1000/e1000e NICs. Crashes occur semi-randomly, starting from 2 days to 2 weeks after bootup. No memory exhaustion can be observed, and CPU load tops at 10% (approx. 5% user + 5% interrupts). My related Fedora bug: https://bugzilla.redhat.com/show_bug.cgi?id=1260391 You can find kernel bug tracebacks there, or via direct links: For 4.1.5: https://bugzilla.redhat.com/attachment.cgi?id=1070710 For 4.1.6: https://bugzilla.redhat.com/attachment.cgi?id=1075750 https://bugzilla.redhat.com/attachment.cgi?id=1075751 I can also directly attach these to this bug if you prefer. I'm currently running 4.1.7 kernel image (stock Fedora) with netconsole enabled - but not enough time has passed yet. Once again, excuse my audacity.
Not sure if related, but i've just got the following warning via netconsole on the last machine running 4.1.6 kernel (attached).
Created attachment 188651 [details] Kernel warning when using 4.1.6
Experienced new crash on 4.1.7 fedora kernel. Crash log is attached. Nothing seems to help.
Created attachment 188681 [details] Crash log when running 4.1.7 kernel