Intermittent network after changing a kvm guest from alma9 to opensuse tumbleweed

[gregrwm]

i want opensuse tumbleweed in a tektonic vps. Tektonic only offers other distributions preinstalled. So first i figured i’d try it here, i installed an alma9 vm on one of our centos8.4 kvm hosts, then did a remote install of opensuse tumbleweed into the vm.

it seemed fine at first but i soon noticed intermittent network response. i’m curious what could cause such intermittent behavior. When changing a kvm guest from alma9 to opensuse tumbleweed is there something else that needs to be accordingly adjusted?

[gregrwm]

Here’s the gory detail of the alma install:

NVM=v151 pw='<hash>' pw=$(printf %q $pw)
cat<<EOI>|/tmp/$NVM.ks
 network --hostname=$NVM
 network --ip=192.168.1.151                       --nameserver=8.8.4.4 --netmask=255.255.248.0 --device=enp1s0 --noipv6 --activate --bootproto=static
 network --ip=192.168.0.151 --gateway=192.168.0.1 --nameserver=8.8.8.8 --netmask=255.255.248.0 --device=enp2s0 --noipv6 --activate --bootproto=static
 timezone America/Chicago --utc
 keyboard --vckeymap=us --xlayouts='us'
 lang en_US.UTF-8
 eula --agreed
 user --name=me --groups=wheel --iscrypted --password=$pw
 repo --name="AppStream" --baseurl=file:///run/install/repo/AppStream
 text
 skipx
 cdrom
 services --disabled="chronyd"
 firstboot --disable
 ignoredisk --only-use=vda
 clearpart --all --initlabel --drives=vda
 part /boot    --fstype="ext4" --ondisk=vda --size=1024
 part pv.01    --fstype="lvmpv" --ondisk=vda --size=2000 --grow
 volgroup vg00 --pesize=4096 pv.01
 logvol /var   --fstype="xfs" --size=4096                      --name=var  --vgname=vg00
 logvol /      --fstype="xfs" --size=5200   --label="root"     --name=root --vgname=vg00
 logvol /home  --fstype="xfs" --size=4096   --label="home"     --name=home --vgname=vg00
 logvol /tmp   --fstype="xfs" --size=4096                      --name=tmp  --vgname=vg00
 logvol swap   --fstype="swap" --size=512                      --name=swap  --vgname=vg00
 logvol /x     --fstype="xfs" --percent=100 --label="removeme" --name=x    --vgname=vg00
 %packages
 @^server-product-environment
 @file-server
 @hardware-monitoring
 @headless-management
 @legacy-unix
 @remote-system-management
 @security-tools
 @system-tools
 tree
 lsscsi
 vim-enhanced
 binutils
 %end
 %post
 echo "net.ipv6.conf.all.disable_ipv6 = 1" >> /etc/sysctl.conf
 setsebool -P domain_kernel_load_modules 1
 umount /x
 lvremove -y /dev/mapper/vg00-x
 sed '/^\/dev\/mapper\/vg00-x/d' /etc/fstab >/tmp/z; mv -f /tmp/z /etc/fstab
 firewall-cmd --permanent --zone=trusted --add-interface=enp1s0
 firewall-cmd --permanent --zone=trusted --add-interface=enp2s0
 firewall --disable
 dnf config-manager --set-enabled crb
 systemctl enable chronyd
 %end
 %addon com_redhat_kdump --disable --reserve-mb='auto'
 %end
EOI
virt-install --name $NVM\
 --nographics --initrd-inject=/tmp/$NVM.ks\
 -x"inst.ks=file:/$NVM.ks ro console=ttyS0,115200n8"\
 --disk path=$path/$NVM-root.qcow2,format=qcow2\
 --location $path/AlmaLinux-9.4-x86_64-dvd.iso\
 --os-type=Linux --os-variant=centos8\
 --network network=br128 --network network=br96\
 --ram=4096 --vcpus=4

The detail of the remote tumbleweed install was all interactive, basically btrfs filesystem, minimal server packageset, and the same network configuration as above.

The nature of the intermittent network response is it seems fine for minutes at a time, then seems dead for a few minutes, then comes back alive again, soon enough that reconnecting isn’t needed, but dead for minutes at a time for no apparent reason, while other VMs are meanwhile having no trouble.

[gregrwm]

@gregrwm

192.168.1.151

sorry, redact error, make that 192.168.32.151