How often do you reboot

[Springs]

I just had to do reboot. I couldn't reach anything inside the network. I have no idea if they were having problems on site. I couldn't reach anything via vpn or port forwards that were behind Untangle. Reboot and all is well. Is there something I can do to help with debugging?

[MichaelTaylor]

I never reboot ... it just runs.

What version are you on?

[Springs]

5.2

[Kumba]

I've noticed a similar behaviour. Processes will slowly creep to high loads, then services will stop working correctly. My particular band-aid to the situation was to set a cron-job to reboot at 5am (1-2 minute "down-time"). Usually this is a sign of a memory leak somewhere or a run-away thread.

As just a side-note, I do notice the problem manifests itself quicker on an intel chipset/CPU then it does on an AMD. Oddly enough, that somewhat makes sense if it's a memory leak/rogue thread. The memory/bus architecture on an AMD can vastly out maneuver the intel design. AMD set-up's I only need to reboot once a week. The Intel set-up's are in the 2-3 day range. for simplicities sake I just set a cron-job to reboot them all nightly. makes life easier

[YeOldeStonecat]

Haven't had to since installing it about 6 months ago on a 70x node network with lots of Exchange server traffic.
Technically it rebooted once...when they lost power for the weekend...so the APC ran dry after a while..no choice.

[Wolf-R1]

I am looking at rebooting as a means for solving a connectivity loss on the inside interface. For some reason every couple of days it looses its connection and refuses to reconnect without a reboot.

[chenzhen]

seems it related to traffic volumn
today I blocked almost all xunlei BT traffic and it haven't hanged for 1 day, will keep testing

[juank]

Never

[dmorris]

rebooting isn't a fix.
All it does is temporarily remove the symptoms and remove evidence that may help you track down the issue.

I know its common in the windows world, but thats because its such a mess that users (myself included) don't even want to bother tracking down the root issues and just want the issue to vanish for the time being.

[Wolf-R1]

No it's not a fix but when you've already put the device into production after testing and testing did not yield this flaw then you have to resort to bandaids and patches to resolve the problem.

In this case it's simple resolution to a seemingly simple problem.