This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

UTM 9.351-3 memory leak

Using UTM 9.351-3, I have a clear memory leak as easily shown by the Weekly Reports I have sent to myself (it *IS* in the daily, but so little that it isn't as obvious as in the Weekly).

Sure, rebooting the UTM temprorarily resolves the high MEM usage, but it always just comes back up.

What can I do to try to figure out why the MEM is leaking like this?



This thread was automatically locked due to age.
Parents
  • There is no memory leak per se. Of course certain releases might have problems with particular daemons but it is not a general problem. I cannot see your graph since you linked it to your gmail account, but if you are just observing the memory usage going up and up till all of it is taken and the system still swapping, that is unfortunately the default behavior. I will say that there is a problem with dual scanning in http proxy. If you only use sophos, the memory usage is way less than if you use avira or dual as your av scanning options.

    Most of us had been fighting with developers in karlsruhe since they didn't tweak the memory management of the kernel much. Now with sophos in charge, the problem has not gone away. You can add some of the memory optimization stuff yourself by using google since this is your home system. I usually add
    echo 1 > /proc/sys/vm/swappiness
    to all my installs. You will have to add the setting after each reboot or you can add it to one of the init scripts like /etc/init.d/beeps to make it stick.

    Generally, if you are using 4gb or more ram, you will not have many problems.

    Hope this helps.

    Edit: Just to clarify, the reason I add it to my init script is because I like keeping all the system changes in one place so that I can quickly refer to all of them in one place. You can of course use traditional sysctl methods if you like.

Reply
  • There is no memory leak per se. Of course certain releases might have problems with particular daemons but it is not a general problem. I cannot see your graph since you linked it to your gmail account, but if you are just observing the memory usage going up and up till all of it is taken and the system still swapping, that is unfortunately the default behavior. I will say that there is a problem with dual scanning in http proxy. If you only use sophos, the memory usage is way less than if you use avira or dual as your av scanning options.

    Most of us had been fighting with developers in karlsruhe since they didn't tweak the memory management of the kernel much. Now with sophos in charge, the problem has not gone away. You can add some of the memory optimization stuff yourself by using google since this is your home system. I usually add
    echo 1 > /proc/sys/vm/swappiness
    to all my installs. You will have to add the setting after each reboot or you can add it to one of the init scripts like /etc/init.d/beeps to make it stick.

    Generally, if you are using 4gb or more ram, you will not have many problems.

    Hope this helps.

    Edit: Just to clarify, the reason I add it to my init script is because I like keeping all the system changes in one place so that I can quickly refer to all of them in one place. You can of course use traditional sysctl methods if you like.

Children
  • Sorry about the image... Let's try this:

    Granted, this shows only increasing MEM usage, so here's one from the next week:

    So the SWAP usage is still steadily increasing. This week's report will be more of the same steady increase (the daily reports show this).
    So if this is expected behaviour, at what point do I enter the realm of what Sophos would "accept" or view as a problem? At what point should the MEM usage taper off at? IS there a max usage I can expect this max out at?