Reducing sys-firewall RAM usage

General Discussion
1

just wondering why i see 1GB usage under xentop, any ways to reduce high RAM usage of sys-firewall?

2

xentop in dom0 only shows how much RAM is allocated, not how much is actually in-use inside the VM. For that, you should run free -h inside the VM. It is usually under 500 MB for sys-firewall.

3

I think an easier way to show the ram consumption is on the little “Q” icon in the “System Tray”.

The other solution might be mirage OS for the sys-firewall:

But I’ve never played with this

(btw, I changed the title to make it a bit more specific so it’s easier for others to find, if they have the same question)

1 Like
4

I have been running sys-firewall with fixed 300MB & 1 core for ages (in qube settings) recently upgraded to fedora-32-minimal template and it’s top looks like:

top - 11:27:42 up 5 days, 10 min,  0 users,  load average: 0.00, 0.00, 0.00
Tasks: 104 total,   1 running, 103 sleeping,   0 stopped,   0 zombie
%Cpu(s):  0.0 us,  6.2 sy,  0.0 ni, 93.8 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
MiB Mem :    265.2 total,     73.9 free,     62.5 used,    128.8 buff/cache
MiB Swap:    512.0 total,    414.0 free,     98.0 used.    193.7 avail Mem 

    PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
   6782 root      20   0   20.5m  10.4m   8.9m S   0.0   3.9   0:00.07 systemd
   6775 root      20   0   18.7m   8.1m   7.2m S   0.0   3.0   0:00.00 systemd-userwor
   6774 root      20   0   18.5m   8.0m   7.1m S   0.0   3.0   0:00.00 systemd-userwor
   6776 root      20   0   18.5m   8.0m   7.1m S   0.0   3.0   0:00.00 systemd-userwor
   6779 root      20   0   76.9m   5.9m   5.0m S   0.0   2.2   0:00.00 qrexec-agent
      1 root      20   0   21.6m   5.8m   4.1m S   0.0   2.2   0:01.07 systemd
    250 root      20   0   49.3m   3.9m   3.5m S   0.0   1.5   0:00.33 systemd-journal
   6803 root      20   0  220.9m   3.7m   3.2m R   0.0   1.4   0:00.00 top
   6788 root      20   0  219.0m   3.5m   3.1m S   0.0   1.3   0:00.01 bash
    632 user      20   0   20.5m   3.1m   2.6m S   0.0   1.2   0:00.10 systemd
   6777 root      20   0   69.8m   2.5m   2.3m S   0.0   1.0   0:00.00 qrexec-agent
    375 root      20   0   18.9m   2.4m   2.1m S   0.0   0.9   0:00.22 systemd-logind
   6783 root      20   0   24.2m   2.2m   0.0m S   0.0   0.8   0:00.00 (sd-pam)
    801 polkitd   20   0 1486.5m   1.2m   0.8m S   0.0   0.4   0:00.11 polkitd
    524 root      20   0   18.4m   1.0m   0.9m S   0.0   0.4   0:02.05 systemd-userdbd
    374 root      20   0   18.9m   1.0m   0.9m S   0.0   0.4   0:00.04 systemd-homed
    760 user      20   0  134.9m   0.8m   0.6m S   0.0   0.3   0:00.14 Xorg

When running lots of qubes it helps to run as many as possible fixed RAM to stop Qubes thrashing mem allocs. Note also a good idea to avoid using X on small system qubes, reducing mem requirement - use qrexec-client from dom0 for any admin, such as above listing. Above suggests 250MB fixed should be adequate. YMMV

[ed] 250MB is just fine, including running xterm - needed to sort out qubes update when it plays up…

2 Likes
5

Thanks, i had the VM at 1GB max and its using 249MB atm.

Q shows me that the VM is hovering at around 1GB of MEM use while inside the VM only 249MB are used.

Reduced to 400MB min and max, as suggested by Eric, lets see how this plays out.
Thanks all.

6

Qubes already handles memory allocation automatically for you. No need to make things harder than they have to be.

2 Likes