Access breaks (SSH, Web, serial console) but firewall still running



  • Since a recent update, my SG-1000 gets itself into a state where I can no longer access the console, via SSH or the web form (504 timeout).  I've been digging through the logs but can't find anything suspicious.  I'm looking for guidance to help identify this problem.  The only thing I added recently was the network traffic graphing package (bandwidthd) which I have now uninstalled.

    Details:

    2.4.0-BETA (arm)
    built on Sun Jul 30 06:26:18 CDT 2017
    FreeBSD 11.0-RELEASE-p11



  • FWIW, after I uninstalled bandwidthd, I haven't seen the lock-out problem repeated.


  • Rebel Alliance Developer Netgate

    Usually a PHP problem would not also cause trouble for SSH, which leads me to believe that the system must be running completely out of RAM.

    What other features and packages do you have enabled? There is not much headroom on SG-1000 so you have to be careful of what is run there since it has no swap space to fall back on.



  • @jimp:

    Usually a PHP problem would not also cause trouble for SSH, which leads me to believe that the system must be running completely out of RAM.

    What other features and packages do you have enabled? There is not much headroom on SG-1000 so you have to be careful of what is run there since it has no swap space to fall back on.

    Interesting.  Nothing shows up in the logs indicating a memory issue and the firewall continues to route traffic at speed.  The packages installed are:

    • AutoConfigBackup

    • Avahi

    • nmap

    Nothing too horrible, right?



  • I'm getting similar issues as well. The only way to recover seems to be power cycle.
    I'm also using bandwidthd as well, so I'll remove it and see if that helps.

    I've also removed RRD_Summary as well.
    So the only packages I have installed are AutoConfigBackup and Service_Watchdog.
    Rebooting now and lets see if it's all ok again



  • @deadmalc:

    I'm getting similar issues as well. The only way to recover seems to be power cycle.
    I'm also using bandwidthd as well, so I'll remove it and see if that helps.

    I've also removed RRD_Summary as well.
    So the only packages I have installed are AutoConfigBackup and Service_Watchdog.
    Rebooting now and lets see if it's all ok again

    I'm glad to hear it wasn't just me.  :P  I'm betting it's a memory issue and the kernel can't free up enough memory to start a shell.



  • Still having the issue, but this time it killed all internet connectivity as well.
    Unfortunately I wasn't at home, so had to power cycle it.
    Not sure this is the same issue as it seems different, but without any access to it at the moment it's impossible to say



  • I noticed the console was non-responsive today. I shutdown from hyper-v and rebooted. It improved for a while then happened again. The only thing I can think of that is related is I was also using winscp to transfer a few files.



  • The console was inoperative this morning. I updated to the latest snapshot. So far, it seems okay, even with winscp connected. Either that may be the result of a change or it's a random coincidence. I'll report back if the problem returns.



  • I guess it was a coincidence. The console was inoperative again this morning.



  • This is definitely a problem. Most often, the console refuses input, but sometimes it also does not display output.



  • I don't think this issue is related to memory. The installation is running in a hyper-v guest with dual processors and 2000 MB of memory. According to the dashboard, only 15% is in use and the disks are almost empty.


Log in to reply