Closed landam closed 7 years ago
@mpastor I guees that the settings should be set from eth0
to eno1
?
The issue with network device should be fixed by 6fe4b95.
The rest of roles finishes successfully.
PLAY RECAP *********************************************************************
gislab-unit-cvut : ok=347 changed=285 unreachable=0 failed=0
Shouldn't be GISLAB_SERVER_NETWORK_DEVICE in Vagrantfile changed from eth1 to eno2? BTW, there are still references in the code to eth0:
providers/gislab-unit/iso/gislab.seed.template:d-i netcfg/choose_interface select eth0
providers/gislab-unit/templates/network/fix-hardware-unit-hang.j2:ethtool -K eth0 tso off
system/roles/client-desktop/templates/initramfs-tools/scripts/init-premount/udhcp.j2:iface=eth0
system/roles/client-desktop/templates/initramfs-tools/scripts/init-premount/udhcp.j2: # Make a DHCP request for interface eth0
system/roles/client-desktop/templates/desktop/conky/conky.conf.j2:${color0}ETH0 ${color1}${downspeed eth0}:${upspeed eth0}
system/group_vars/all:# 'eth0' and setting default gateway via 'eth1' device.
system/group_vars/all:# Removing original default gateway: ip route del default via 10.0.2.2 dev eth0
In summary:
@imincik @mpastor Make sense to you?
@landam , network devices names can vary in different systems. It is very complicated, but I have some ideas.
For sure, the network names should be replaced by ansible network device variables, they shouldn't be hardcoded.
In "client-desktop" we can even implement predictable network names, because we have full control on that system, but with server we must probably finish with some better solution.
@imincik , can you provide some details about your idea? Thanks!
All issues seems to be solved.
After GIS.lab initialization (see #496) I run
which failed with