1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
|
# Testbed Setup
## Introduction
This directoctory contains the *high-level* process to set up a hardware
machine as a CSIT testbed, either for use as a physical testbed host or
as a VIRL server.
Code in this directory is NOT executed as part of a regular CSIT test case
but is stored here merely for archiving and documentation purposes.
## Setting up a hardware host
Documentation below is just bullet points and assumes and understanding
of PXE boot and ansible.
This process is specific for LF lab, and both examples given here as
well as associated code, are based on the assumption that they are run
in LF environment. If run elsewhere, changes will be required to IP addresses
and other parameters.
The process below assumes that there is a host used for boostrapping (referred
to as "PXE boostrap server" below), and that the directory containig this README
is available on the PXE bootstrap server in ~testuser/host-setup.
### Prepare the PXE bootstrap server when there is no http server
- `sudo apt-get install isc-dhcp-server tftpd-hpa nginx-light ansible`
- edit dhcpd.conf and place it to /etc/dhcp/
- `sudo cp dhcpd.cfg /etc/dhcp/`
- `sudo service isc-dhcp-server restart`
- `cd ~testuser/host-setup`
- `wget 'http://releases.ubuntu.com/16.04.2/ubuntu-16.04.2-server-amd64.iso'`
- `sudo mkdir /mnt/cdrom`
- `sudo mount -o loop ubuntu-16.04.2-server-amd64.iso /mnt/cdrom/`
- `sudo cp -r /mnt/cdrom/install/netboot/* /var/lib/tftpboot/`
- figure out where nginx will look for files on the filesystem when
responding to HTTP requests. The configuration is in one of the
files in /etc/nginx/conf.d/, /etc/nginx/sites-enabled/ or in
/etc/nginx/nginx.conf under section server/root. Save the path to NGINX_ROOT
- `sudo mkdir -p ${NGINX_ROOT}/download/ubuntu`
- `sudo cp -r /mnt/cdrom/* ${NGINX_ROOT}/download/ubuntu/`
- `sudo cp /mnt/cdrom/ubuntu/isolinux/ldlinux.c32 /var/lib/tftpboot`
- `sudo cp /mnt/cdrom/ubuntu/isolinux/libcom32.c32 /var/lib/tftpboot`
- `sudo cp /mnt/cdrom/ubuntu/isolinux/libutil.c32 /var/lib/tftpboot`
- `sudo cp /mnt/cdrom/ubuntu/isolinux/chain.c32 /var/lib/tftpboot`
- `sudo umount /mnt/cdrom`
- edit ks.cfg and replace IP address with that of your PXE bootstrap server
- `sudo cp ks.cfg ${NGINX_ROOT}/download/ks.cfg`
- edit boot-screens_txt.cfg and replace IP address with that of your PXE bootstrap server
- `sudo cp boot-screens_txt.cfg /var/lib/tftpboot/ubuntu-installer/amd64/boot-screens/txt.cfg`
- `sudo cp syslinux.cfg /var/lib/tftpboot/ubuntu-installer/amd64/boot-screens/syslinux.cfg`
### PREFERED: Prepare the PXE bootstrap server when an http server is already configured
- `sudo apt-get install isc-dhcp-server tftpd-hpa ansible`
- edit dhcpd.conf and place it to /etc/dhcp/
- `sudo cp dhcpd.cfg /etc/dhcp/`
- `sudo service isc-dhcp-server restart`
- `cd ~testuser/host-setup`
- `wget 'http://releases.ubuntu.com/16.04.2/ubuntu-16.04.2-server-amd64.iso'`
- `sudo mkdir /mnt/cdrom`
- `sudo mount -o loop ubuntu-16.04.1-server-amd64.iso /mnt/cdrom/`
- `sudo cp -r /mnt/cdrom/install/netboot/* /var/lib/tftpboot/`
- `sudo mkdir /var/www/download/ubuntu`
- `sudo cp -r /mnt/cdrom/* /var/www/download/ubuntu/`
- `sudo cp /mnt/cdrom/ubuntu/isolinux/ldlinux.c32 /var/lib/tftpboot`
- `sudo cp /mnt/cdrom/ubuntu/isolinux/libcom32.c32 /var/lib/tftpboot`
- `sudo cp /mnt/cdrom/ubuntu/isolinux/libutil.c32 /var/lib/tftpboot`
- `sudo cp /mnt/cdrom/ubuntu/isolinux/chain.c32 /var/lib/tftpboot`
- `sudo umount /mnt/cdrom`
- edit ks.cfg and replace IP address with that of your PXE bootstrap server and subdir in /var/www (in this case /download)
- `sudo cp ks.cfg /var/www/download/ks.cfg`
- edit boot-screens_txt.cfg and replace IP address with that of your PXE bootstrap server and subdir in /var/www (in this case /download)
- `sudo cp boot-screens_txt.cfg /var/lib/tftpboot/ubuntu-installer/amd64/boot-screens/txt.cfg`
- `sudo cp syslinux.cfg /var/lib/tftpboot/ubuntu-installer/amd64/boot-screens/syslinux.cfg`
### New testbed host - manual preparation
- set CIMC address
- set CIMC username, password and hostname
- set IPMI address
- set IPMI username, password and hostname
### Bootstrap the host
Optional: From PXE boostrap server in case of installing Haswell
- `cd resources/tools/testbed-setup/cimc`
- Initialize args.ip: Power-Off, reset BIOS defaults, Enable console redir, get LOM MAC addr
- `./cimc.py -u admin -p Cisco1234 $CIMC_ADDRESS -d -i`
- Adjust BIOS settings
- `./cimc.py -u admin -p Cisco1234 $CIMC_ADDRESS -d -s '<biosVfIntelHyperThreadingTech rn="Intel-HyperThreading-Tech" vpIntelHyperThreadingTech="disabled" />' -s '<biosVfEnhancedIntelSpeedStepTech rn="Enhanced-Intel-SpeedStep-Tech" vpEnhancedIntelSpeedStepTech="disabled" />' -s '<biosVfIntelTurboBoostTech rn="Intel-Turbo-Boost-Tech" vpIntelTurboBoostTech="disabled" />'`
- add MAC address to DHCP (/etc/dhcp/dhcpd.conf)
- Reboot server with boot from PXE (restart immediately)
- `./cimc.py -u admin -p Cisco1234 $CIMC_ADDRESS -d -pxe`
Optional: If RAID is not created on Haswells. Execute while Ubuntu install is running
- create RAID array. Reboot if needed.
- `./cimc.py -u admin -p Cisco1234 $CIMC_ADDRESS -d --wipe`
- `./cimc.py -u admin -p Cisco1234 $CIMC_ADDRESS -d -r -rl 1 -rs <disk size> -rd '[1,2]'`
Alternatively, create the RAID array manually.
- Set the next boot from HDD (without restart)
- `./cimc.py -u admin -p Cisco1234 $CIMC_ADDRESS -d -hdd`
Optional: If installing Skylake machine, connect to IPMI and boot from PXE via F12
When installation is finished:
- Copy ssh keys for no pass access: `ssh-copy-id 10.30.51.x`
- Clone CSIT actual repo: `git clone https://gerrit.fd.io/r/csit`
- Go to ansible directory: `cd csit/resources/tools/testbed-setup/ansible`
- Edit production file and uncomment servers that are supposed to be installed: `ansible-playbook --ask-become-pass --inventory production site.yaml --list-hosts`
- Run ansible on selected hosts: `ansible-playbook --ask-become-pass --inventory production site.yaml`
For non-VIRL hosts, stop here.
### VIRL installation
After the host has rebooted:
- `ansible-playbook 02-virl-bootstrap.yaml`
- ssh to host
- `sudo -s`
- `cd virl-bootstrap`
- `./virl-bootstrap-wrapper`
This command will error out when run the first time, as the VIRL host is not yet licensed.
Make sure we contact all three VIRL SALT masters:
- `for a in 1 2 4 ; do sudo salt-call --master us-${a}.virl.info test.ping ; done`
- Contact the VIRL team, provide the hostname and domain (linuxfoundation.org), and ask them
to accept the key
- After the key has been accepted, verify that connectivity with the SALT master is now OK:
`for a in 1 2 4 ; do sudo salt-call --master us-${a}.virl.info test.ping ; done`
- `./virl-bootstrap-wrapper`
- `reboot`
After reboot, ssh to host again
- as VIRL user, NOT AS ROOT:
- `vinstall all`
- `sudo reboot`
After reboot, ssh to host again
- as VIRL user:
- `sudo salt-call state.sls virl.routervms.all`
- `sudo salt-call state.sls virl.vmm.vmmall`
Back on the PXE bootstrap server:
- obtain the current server disk image and place it into
`files/virl-server-image/` as `server.qcow2`
TO-DO: Need to find a place to store this image
- `ansible-playbook 03-virl-post-install.yaml`
- Run the following command ONLY ONCE. Otherwise it will create
duplicates of the VIRL disk image:
`ansible-playbook 04-disk-image.yaml`
The VIRL host should now be operational. Test, and when ready, create a ~jenkins-in/status file with the appropriate status.
|
