xbps icon indicating copy to clipboard operation
xbps copied to clipboard

Void with 512 megs of RAM (VM) - xbps sync causes OOMkill

Open Vaelatern opened this issue 8 months ago • 3 comments

In the 2024 ISO, this wasn't a problem. Now I have a packer build as follows:

└» cat void-autoinstall.pkr.hcl
source "qemu" "void-musl" {
  iso_url          = "https://repo-fastly.voidlinux.org/live/current/void-live-x86_64-musl-20250202-base.iso"
  iso_checksum     = "sha256:0c46bf13a7a019c1fec0b2969201aecf09ff9a7646ef749e04ea63ead0c86404"
  output_directory = "output-void-musl"
  shutdown_command = "echo 'voidlinux' | sudo -S shutdown -P now"
  disk_size        = "25G"
  format           = "qcow2"
  accelerator      = "kvm"
  headless         = true
  http_directory   = "http"

  ssh_username = "admin"
  ssh_password = "voidlinux"
  ssh_timeout  = "90m" # How long we give the system to get installed

  boot_wait = "2s"
  boot_command = [
    "<tab><wait>",
    "auto autourl=http://{{.HTTPIP}}:{{.HTTPPort}}/basic-void.cfg",
    "<enter>"
  ]
}

build {
  source "source.qemu.void-musl" {
    vm_name = "void-nova"
  }
}

and the config I load:

└» cat http/basic-void.cfg
#!/bin/sh

export username=admin
export password=voidlinux

export XBPS_ARCH=x86_64-musl
export xbpsrepository="https://repo-fastly.voidlinux.org/current/musl"

export timezone="America/New_York"

export pkgs="docker git neovim temporal-cli nomad uv"

export end_action=func
end_function() {
        printf "Linking default services\n"
        ln -s /etc/sv/docker "${target}/etc/runit/runsvdir/default/docker"
        ln -s /etc/sv/dhcpcd "${target}/etc/runit/runsvdir/default/dhcpcd"
        ln -s /etc/sv/sshd "${target}/etc/runit/runsvdir/default/sshd"
        ln -s /etc/sv/agetty-ttyS0 "${target}/etc/runit/runsvdir/default/agetty-ttyS0"
        echo "***"


        echo "***"
        printf "Let's continue with our life.\n"
        sync
        reboot -f
}

# Skip boot partition
VAI_partition_disk() {
        sfdisk "${disk}" <<EOF
,${swapsize},S
;
EOF
}
VAI_format_disk() {
        mkfs.ext4 -F "${disk}2"
        mkswap -f "${disk}1"
}
export swapsize="1G"

VAI_mount_target() {
        mkdir -p "${target}"
        mount "${disk}2" "${target}"
}

VAI_configure_fstab() {
        uuid_root="$(blkid -s UUID -o value "${disk}2")"
        uuid_swap="$(blkid -s UUID -o value "${disk}1")"
        echo "UUID=$uuid_root / ext4 defaults,errors=remount-ro,noatime 0 1" >> "${target}/etc/fstab"
        echo "UUID=$uuid_swap swap swap defaults 0 0" >> "${target}/etc/fstab"
}

# Removable when this is closed
VAI_install_base_system() {
    # Install a base system
    XBPS_ARCH="${XBPS_ARCH}" xbps-install -Sy -R "${xbpsrepository}" -r /mnt base-system grub

    # Install additional packages
    if [  -n "${pkgs}" ] ; then
        # shellcheck disable=SC2086
        XBPS_ARCH="${XBPS_ARCH}" xbps-install -Sy -R "${xbpsrepository}" -r /mnt ${pkgs}
    fi
}

(yes most of this can be removed in the newest mklive but the newest mklive is newer than the ISO)

The point is this image ends up in trouble because xbps-install triggers an OOMkill reliably. I believe xbps is allocating a lot of memory where it didn't always, and the system has a 512 meg limit. If I increase the system RAM from the default (512) to 648 megs, the install works flawlessly.

Vaelatern avatar Jun 27 '25 02:06 Vaelatern

Might be fixed now, I reduced a bunch of temporary allocations, most memory is basically spent on keeping repository indexes and package database in memory since they are not the most compact format there is. But it never actually allocated that much memory in total, I think its just that it temporarily allocated a bunch of stuff.

https://github.com/void-linux/xbps/pull/636

Duncaen avatar Aug 02 '25 18:08 Duncaen

I'll test again

Vaelatern avatar Aug 10 '25 10:08 Vaelatern

Very impressive!

I had to set +e around the tar command in this script because tar would OOM. I had to remove the tarball and excess files because if I didn't the system would OOM. But xbps-install with 0.60.4_1 was able to install a whole system with the default 512 megs provided by qemu

└» find
.
./output-void-musl
./output-void-musl/void-nova
./http
./http/basic-void.cfg
./void-autoinstall.pkr.hcl
└» cat ./void-autoinstall.pkr.hcl
source "qemu" "void-musl" {
  iso_url          = "https://repo-fastly.voidlinux.org/live/current/void-live-x86_64-musl-20250202-base.iso"
  iso_checksum     = "sha256:0c46bf13a7a019c1fec0b2969201aecf09ff9a7646ef749e04ea63ead0c86404"
  output_directory = "output-void-musl"
  shutdown_command = "echo 'voidlinux' | sudo -S shutdown -P now"
  disk_size        = "25G"
  format           = "qcow2"
  accelerator      = "kvm"
  headless         = true
  http_directory   = "http"

  ssh_username = "admin"
  ssh_password = "voidlinux"
  ssh_timeout  = "90m" # How long we give the system to get installed

  boot_wait = "2s"
  boot_command = [
    "<tab><wait>",
    "auto autourl=http://{{.HTTPIP}}:{{.HTTPPort}}/basic-void.cfg",
    "<enter>"
  ]
}

build {
  source "source.qemu.void-musl" {
    vm_name = "void-nova"
  }
}
└» cat ./http/basic-void.cfg
#!/bin/sh

export username=admin
export password=voidlinux

export XBPS_ARCH=x86_64-musl
export xbpsrepository="https://repo-fastly.voidlinux.org/current/musl"

export timezone="America/New_York"

export pkgs="docker git neovim temporal-cli nomad uv"

export end_action=func
end_function() {
        printf "Linking default services\n"
        ln -s /etc/sv/docker "${target}/etc/runit/runsvdir/default/docker"
        ln -s /etc/sv/dhcpcd "${target}/etc/runit/runsvdir/default/dhcpcd"
        ln -s /etc/sv/sshd "${target}/etc/runit/runsvdir/default/sshd"
        ln -s /etc/sv/agetty-ttyS0 "${target}/etc/runit/runsvdir/default/agetty-ttyS0"
        echo "***"


        echo "***"
        printf "Let's continue with our life.\n"
        sync
        reboot -f
}

# Skip boot partition
VAI_partition_disk() {
        sfdisk "${disk}" <<EOF
,${swapsize},S
;
EOF
}
VAI_format_disk() {
        mkfs.ext4 -F "${disk}2"
        mkswap -f "${disk}1"
}
export swapsize="1G"

VAI_mount_target() {
        mkdir -p "${target}"
        mount "${disk}2" "${target}"
}

VAI_configure_fstab() {
        uuid_root="$(blkid -s UUID -o value "${disk}2")"
        uuid_swap="$(blkid -s UUID -o value "${disk}1")"
        echo "UUID=$uuid_root / ext4 defaults,errors=remount-ro,noatime 0 1" >> "${target}/etc/fstab"
        echo "UUID=$uuid_swap swap swap defaults 0 0" >> "${target}/etc/fstab"
}

# Removable when this is closed
VAI_install_base_system() {
    cd /var/
    xbps-uhelper fetch https://repo-default.voidlinux.org/static/xbps-static-static-0.60.4_1.x86_64-musl.tar.xz
    set +e
    tar -xvf xbps-static*
    set -e
    mv usr/bin/xbps-install.static /bin/xbps-install
    rm -rf usr xbps-static*
    xbps-install -V
    # Install a base system
    XBPS_ARCH="${XBPS_ARCH}" xbps-install -Sy -R "${xbpsrepository}" -r /mnt base-system grub

    # Install additional packages
    if [  -n "${pkgs}" ] ; then
        # shellcheck disable=SC2086
        XBPS_ARCH="${XBPS_ARCH}" xbps-install -y -R "${xbpsrepository}" -r /mnt ${pkgs}
    fi
}

If I add void-repo-nonfree to the mix, the additional repository does seem to scale past the remaining memory. Might be able to squeeze by if I'd been using an image with xbps-install not being replaced.

Vaelatern avatar Nov 25 '25 11:11 Vaelatern