Elastix PBX Memory question



  • Doing some clarifications on memory. and rebooting a system.

    The GUI gives you the option to reboot the system. This is not the same reboot option as from the CLI.

    If you perform a GUI reboot, does it free up memory the same way you would do a CLI reboot?

    Tagging:
    @scottalanmiller @JaredBusch



  • The GUI reboot should be identical to the CLI reboot. A reboot is a reboot.



  • If a reboot is needed to clear memory, that indicates that something is wrong. You should look to see what is using the memory before rebooting.



  • System had been running 144 days 7 hours 42 minutes when I accessed it. Memory usage was about 256MB for ram, 509MB for the swap file.

    Problem turned out to be an issue with the provider. They self rebooted using the GUI once the provider came back up.

    Normal operation as of now,.. Just curious if there was a difference between the two modes of reboot.



  • @g.jacobse said:

    System had been running 144 days 7 hours 42 minutes when I accessed it. Memory usage was about 256MB for ram, 509MB for the swap file.

    Nothing wrong with that. Rebooting to keep uptime down is good, but that memory is probably just fine. Did you look at a top sorted by memory usage or look at sysstat to see if there was any active paging activity? Having 256MB used and 509MB for swap doesn't indicate any memory issue or exhaustion on its own.

    What was the output of free -m?



  • @g.jacobse said:

    Normal operation as of now,.. Just curious if there was a difference between the two modes of reboot.

    Nope, only one system signal that causes a reboot. So anything that triggers that will be identical.



  • You should be able to look at a sar report and have a pretty good idea of if there were memory issues. But I know the system in question and that doesn't sound like anything wrong. It was very unlikely that it was paging.



  • People always like to claim these high uptimes and I always wonder if that means they don't patch their systems. Ours rarely go longer than 7-14 days without a reboot.



  • @thecreativeone91 said:

    People always like to claim these high uptimes and I always wonder if that means they don't patch their systems. Ours rarely go longer than 7-14 days without a reboot.

    I like to call this one "admitting" a long uptime. I agree, ideally reboot every seven days (just because an "every other" weekend greenzone is very awkward to track and remember.)

    Problem with PBXs is not many businesses want a weekly downtime, even of a minute or two. Most patching can be done without downtime, only kernels typically cause a need for downtime and that is much less common.



  • @thecreativeone91 said:

    People always like to claim these high uptimes and I always wonder if that means they don't patch their systems. Ours rarely go longer than 7-14 days without a reboot.

    I've not following the patch / revision cycle. But is your 'low up time' due to that?



  • @g.jacobse said:

    I've not following the patch / revision cycle. But is your 'low up time' due to that?

    How do you patch reliably without rebooting to verify it or, in most cases, how do you patch at all without rebooting?



  • He is referring to this philosophy, above and beyond direct patching concerns:

    Why We Reboot Servers


Log in to reply