PfSense Crashing Every Day



  • I've tried two different boxes, and the same thing happens on both, i've also installed pfsense fresh.(Currently on 2.2.3 and it just crash last night) I have a feeling something in the config is causing it. I submit the crash report 66.225.80.5 at 10:47am EST. Any help, even a nudge in the right direction would help.

    Thanks,



  • With absolutely no information given, I don't know what you expect people to do to try and help you.



  • Does the crash report have stuff about PHP error/warning?
    If so, then post the details here (error message and file/line number).

    Otherwise, is it an operating system crash? Does the whole system reboot?

    What is your feeling based on?
    What did you do in the config that makes you think it has a problem?

    Information will help us help you. Only a few people have access to the crash reports, and I they will not necessarily be looking here every hour.



  • I added this post based on this KB article https://doc.pfsense.org/index.php/Obtaining_Panic_Information_for_Developers.

    Anyways i think it's a config problem because I have had the problem with completely separate servers The only thing that was the same between them was the config. It's crashed with 2.2.1 2.2.2 and now 2.2.3.

    So far I've tried

    Fresh Install
    2 separate servers

    It seems to be load related somewhat, because i can run the config fine on bench for weeks without crashing. (The server that is crashing now, ran fine for 2 weeks on my desk before i put it into production, within a couple of days it started crashing). The whole system reboots.

    I have another pfsense with like hardware running a very different config that never crashes.

    Attached is the crash report i submitted.

    pfsense_crash_report.txt



  • Also not sure if this is related, but this process increases in memory usage until all free memory is gone, it's replaced with inactive memory and reading about Freebsd memory usage it didn't seem to be an issue. None of my other boxes do this, but then again i have 100+ ipsec tunnels on this box and the others only have a couplel at most.

    ps -aux
    USER      PID  %CPU %MEM    VSZ    RSS    TT  STAT STARTED       TIME COMMAND
    
    root      46797  0.0     24.5   653908 440900  -  Ss    6:16AM     3:33.06 /usr/local/libexec/ipsec/charon --use-syslog
    


  • Fatal trap 12: page fault while in kernel mode
    

    That certainly is not supposed to happen! Pages used by kernel mode code are supposed to be already memory resident.
    Not having any FreeBSD crash-dump analysis knowledge I will leave it to others to comment…



  • I also noticed this in the crash report. Around 6:15am is when the crash happend.

    <6>pid 19814 (charon), uid 0: exited on signal 6 (core dumped)
    <118>Jun 26 06:14:56 ipsec_starter[19615]: routing 'con73' failed
    <118>Jun 26 06:14:56 ipsec_starter[19615]: 
    <118>Jun 26 06:14:56 ipsec_starter[19615]: routing 'con1000' failed
    <118>Jun 26 06:14:56 ipsec_starter[19615]: 
    <118>Jun 26 06:14:56 ipsec_starter[19615]: routing 'con1001' failed
    

    ….....



  • Watch, Charon isn't flagging memory pages to not be paged out then the kernel attempts to access said memory locations.


Log in to reply