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

High memory swapping since update to Sophos UTM 9.714-4

Hi all,

before the update to Sophos UTM 9.714-4 I had 10% swap usage in maximum. Since the update it increased day by day until yesterday to about 75%. Then I shut down the HA-Cluster (active/passive) and started the two nodes again. The swap usage increased again to 20%.

There was no change of the configuration. There was only the update to Sophos UTM 9.714-4.

Can anyone confirm this behavior of the newest release of the Sophos UTM? Is there any idea to solve this?

Thank you.



This thread was automatically locked due to age.
Parents
  • Hi all,

    this behavoir is also described in this thread: High httpproxy memory usage with 9.713 and 9.714 - General Discussion - UTM Firewall - Sophos Community.

    The only difference is that until release 9.713 I didn't have a high swap usage. It increased with 9.714-4.

    Kind Regards

    TheExpert

  • Please see the answer I posted on the other thread - in short, "Swap usage" is "Swap space that has been written to at some point" and not "Swap activity" or even "Swap space containing pages of memory that might be required immediately". If a system ever needs to use Swap, the 'Swap usage' metric will never go back down and will continue to increase each time swap is required, even briefly.

  • Thank you. But in the former releases of Sophos UTM the swap usage was constant and when it increased it also decreased. Now it's only increasing.

    Kind Regards

    TheExpert

  • If you have the bare minimum memory installed on UTM, especially with 64-bit mode changing in 9.713, you will see the swap increase exponentially. If this is a VM, I would allocate more memory to it. 

    This is the issue I have with UTM (and XG) - it's great that things are finally going 64-bit.  The problem is the hardware originally allocated (either original hardware or VM when built) to them were not. UTM isn't a big deal, just add more memory. I have 16GB for UTM, 0% swap usage.  Home use XG users though are limited and (honestly) a bad design decision.  Out of the box with no configuration you are using 60% memory, and it just gets worse each time you add a rule, change configuration, etc. 

    OPNSense 64-bit | Intel Xeon 4-core v3 1225 3.20Ghz
    16GB Memory | 500GB SSD HDD | ATT Fiber 1GB
    (Former Sophos UTM Veteran, Former XG Rookie)

Reply
  • If you have the bare minimum memory installed on UTM, especially with 64-bit mode changing in 9.713, you will see the swap increase exponentially. If this is a VM, I would allocate more memory to it. 

    This is the issue I have with UTM (and XG) - it's great that things are finally going 64-bit.  The problem is the hardware originally allocated (either original hardware or VM when built) to them were not. UTM isn't a big deal, just add more memory. I have 16GB for UTM, 0% swap usage.  Home use XG users though are limited and (honestly) a bad design decision.  Out of the box with no configuration you are using 60% memory, and it just gets worse each time you add a rule, change configuration, etc. 

    OPNSense 64-bit | Intel Xeon 4-core v3 1225 3.20Ghz
    16GB Memory | 500GB SSD HDD | ATT Fiber 1GB
    (Former Sophos UTM Veteran, Former XG Rookie)

Children
  • Hi all,

    after a new restart of the Sophos UTM cluster nodes because of updating the ESXi 8.0 hosts the swap usage keeps to be around 22%. This is OK at the moment. If this will increase again I will add more memory.

    As I wrote before this "issue" started for me with 9.714 and not with 9.713.

    How can I check If the nodes are installed in 64 bit mode vs. 64 bit compatibility mode? I'm sure I installed them in 64 bit mode but this was a long time ago.

    Kind Regards

    TheExpert