github.com/enmand/kubernetes@v1.2.0-alpha.0/docs/getting-started-guides/coreos/bare_metal_offline.md (about) 1 <!-- BEGIN MUNGE: UNVERSIONED_WARNING --> 2 3 <!-- BEGIN STRIP_FOR_RELEASE --> 4 5 <img src="http://kubernetes.io/img/warning.png" alt="WARNING" 6 width="25" height="25"> 7 <img src="http://kubernetes.io/img/warning.png" alt="WARNING" 8 width="25" height="25"> 9 <img src="http://kubernetes.io/img/warning.png" alt="WARNING" 10 width="25" height="25"> 11 <img src="http://kubernetes.io/img/warning.png" alt="WARNING" 12 width="25" height="25"> 13 <img src="http://kubernetes.io/img/warning.png" alt="WARNING" 14 width="25" height="25"> 15 16 <h2>PLEASE NOTE: This document applies to the HEAD of the source tree</h2> 17 18 If you are using a released version of Kubernetes, you should 19 refer to the docs that go with that version. 20 21 <strong> 22 The latest 1.0.x release of this document can be found 23 [here](http://releases.k8s.io/release-1.0/docs/getting-started-guides/coreos/bare_metal_offline.md). 24 25 Documentation for other releases can be found at 26 [releases.k8s.io](http://releases.k8s.io). 27 </strong> 28 -- 29 30 <!-- END STRIP_FOR_RELEASE --> 31 32 <!-- END MUNGE: UNVERSIONED_WARNING --> 33 Bare Metal CoreOS with Kubernetes (OFFLINE) 34 ------------------------------------------ 35 Deploy a CoreOS running Kubernetes environment. This particular guild is made to help those in an OFFLINE system, wither for testing a POC before the real deal, or you are restricted to be totally offline for your applications. 36 37 **Table of Contents** 38 39 - [Prerequisites](#prerequisites) 40 - [High Level Design](#high-level-design) 41 - [This Guides variables](#this-guides-variables) 42 - [Setup PXELINUX CentOS](#setup-pxelinux-centos) 43 - [Adding CoreOS to PXE](#adding-coreos-to-pxe) 44 - [DHCP configuration](#dhcp-configuration) 45 - [Kubernetes](#kubernetes) 46 - [Cloud Configs](#cloud-configs) 47 - [master.yml](#masteryml) 48 - [node.yml](#nodeyml) 49 - [New pxelinux.cfg file](#new-pxelinuxcfg-file) 50 - [Specify the pxelinux targets](#specify-the-pxelinux-targets) 51 - [Creating test pod](#creating-test-pod) 52 - [Helping commands for debugging](#helping-commands-for-debugging) 53 54 55 ## Prerequisites 56 57 1. Installed *CentOS 6* for PXE server 58 2. At least two bare metal nodes to work with 59 60 ## High Level Design 61 62 1. Manage the tftp directory 63 * /tftpboot/(coreos)(centos)(RHEL) 64 * /tftpboot/pxelinux.0/(MAC) -> linked to Linux image config file 65 2. Update per install the link for pxelinux 66 3. Update the DHCP config to reflect the host needing deployment 67 4. Setup nodes to deploy CoreOS creating a etcd cluster. 68 5. Have no access to the public [etcd discovery tool](https://discovery.etcd.io/). 69 6. Installing the CoreOS slaves to become Kubernetes nodes. 70 71 ## This Guides variables 72 73 | Node Description | MAC | IP | 74 | :---------------------------- | :---------------: | :---------: | 75 | CoreOS/etcd/Kubernetes Master | d0:00:67:13:0d:00 | 10.20.30.40 | 76 | CoreOS Slave 1 | d0:00:67:13:0d:01 | 10.20.30.41 | 77 | CoreOS Slave 2 | d0:00:67:13:0d:02 | 10.20.30.42 | 78 79 80 ## Setup PXELINUX CentOS 81 82 To setup CentOS PXELINUX environment there is a complete [guide here](http://docs.fedoraproject.org/en-US/Fedora/7/html/Installation_Guide/ap-pxe-server.html). This section is the abbreviated version. 83 84 1. Install packages needed on CentOS 85 86 sudo yum install tftp-server dhcp syslinux 87 88 2. `vi /etc/xinetd.d/tftp` to enable tftp service and change disable to 'no' 89 disable = no 90 91 3. Copy over the syslinux images we will need. 92 93 su - 94 mkdir -p /tftpboot 95 cd /tftpboot 96 cp /usr/share/syslinux/pxelinux.0 /tftpboot 97 cp /usr/share/syslinux/menu.c32 /tftpboot 98 cp /usr/share/syslinux/memdisk /tftpboot 99 cp /usr/share/syslinux/mboot.c32 /tftpboot 100 cp /usr/share/syslinux/chain.c32 /tftpboot 101 102 /sbin/service dhcpd start 103 /sbin/service xinetd start 104 /sbin/chkconfig tftp on 105 106 4. Setup default boot menu 107 108 mkdir /tftpboot/pxelinux.cfg 109 touch /tftpboot/pxelinux.cfg/default 110 111 5. Edit the menu `vi /tftpboot/pxelinux.cfg/default` 112 113 default menu.c32 114 prompt 0 115 timeout 15 116 ONTIMEOUT local 117 display boot.msg 118 119 MENU TITLE Main Menu 120 121 LABEL local 122 MENU LABEL Boot local hard drive 123 LOCALBOOT 0 124 125 Now you should have a working PXELINUX setup to image CoreOS nodes. You can verify the services by using VirtualBox locally or with bare metal servers. 126 127 ## Adding CoreOS to PXE 128 129 This section describes how to setup the CoreOS images to live alongside a pre-existing PXELINUX environment. 130 131 1. Find or create the TFTP root directory that everything will be based off of. 132 * For this document we will assume `/tftpboot/` is our root directory. 133 2. Once we know and have our tftp root directory we will create a new directory structure for our CoreOS images. 134 3. Download the CoreOS PXE files provided by the CoreOS team. 135 136 MY_TFTPROOT_DIR=/tftpboot 137 mkdir -p $MY_TFTPROOT_DIR/images/coreos/ 138 cd $MY_TFTPROOT_DIR/images/coreos/ 139 wget http://stable.release.core-os.net/amd64-usr/current/coreos_production_pxe.vmlinuz 140 wget http://stable.release.core-os.net/amd64-usr/current/coreos_production_pxe.vmlinuz.sig 141 wget http://stable.release.core-os.net/amd64-usr/current/coreos_production_pxe_image.cpio.gz 142 wget http://stable.release.core-os.net/amd64-usr/current/coreos_production_pxe_image.cpio.gz.sig 143 gpg --verify coreos_production_pxe.vmlinuz.sig 144 gpg --verify coreos_production_pxe_image.cpio.gz.sig 145 146 4. Edit the menu `vi /tftpboot/pxelinux.cfg/default` again 147 148 default menu.c32 149 prompt 0 150 timeout 300 151 ONTIMEOUT local 152 display boot.msg 153 154 MENU TITLE Main Menu 155 156 LABEL local 157 MENU LABEL Boot local hard drive 158 LOCALBOOT 0 159 160 MENU BEGIN CoreOS Menu 161 162 LABEL coreos-master 163 MENU LABEL CoreOS Master 164 KERNEL images/coreos/coreos_production_pxe.vmlinuz 165 APPEND initrd=images/coreos/coreos_production_pxe_image.cpio.gz cloud-config-url=http://<xxx.xxx.xxx.xxx>/pxe-cloud-config-single-master.yml 166 167 LABEL coreos-slave 168 MENU LABEL CoreOS Slave 169 KERNEL images/coreos/coreos_production_pxe.vmlinuz 170 APPEND initrd=images/coreos/coreos_production_pxe_image.cpio.gz cloud-config-url=http://<xxx.xxx.xxx.xxx>/pxe-cloud-config-slave.yml 171 MENU END 172 173 This configuration file will now boot from local drive but have the option to PXE image CoreOS. 174 175 ## DHCP configuration 176 177 This section covers configuring the DHCP server to hand out our new images. In this case we are assuming that there are other servers that will boot alongside other images. 178 179 1. Add the `filename` to the _host_ or _subnet_ sections. 180 181 filename "/tftpboot/pxelinux.0"; 182 183 2. At this point we want to make pxelinux configuration files that will be the templates for the different CoreOS deployments. 184 185 subnet 10.20.30.0 netmask 255.255.255.0 { 186 next-server 10.20.30.242; 187 option broadcast-address 10.20.30.255; 188 filename "<other default image>"; 189 190 ... 191 # http://www.syslinux.org/wiki/index.php/PXELINUX 192 host core_os_master { 193 hardware ethernet d0:00:67:13:0d:00; 194 option routers 10.20.30.1; 195 fixed-address 10.20.30.40; 196 option domain-name-servers 10.20.30.242; 197 filename "/pxelinux.0"; 198 } 199 host core_os_slave { 200 hardware ethernet d0:00:67:13:0d:01; 201 option routers 10.20.30.1; 202 fixed-address 10.20.30.41; 203 option domain-name-servers 10.20.30.242; 204 filename "/pxelinux.0"; 205 } 206 host core_os_slave2 { 207 hardware ethernet d0:00:67:13:0d:02; 208 option routers 10.20.30.1; 209 fixed-address 10.20.30.42; 210 option domain-name-servers 10.20.30.242; 211 filename "/pxelinux.0"; 212 } 213 ... 214 } 215 216 We will be specifying the node configuration later in the guide. 217 218 ## Kubernetes 219 220 To deploy our configuration we need to create an `etcd` master. To do so we want to pxe CoreOS with a specific cloud-config.yml. There are two options we have here. 221 1. Is to template the cloud config file and programmatically create new static configs for different cluster setups. 222 2. Have a service discovery protocol running in our stack to do auto discovery. 223 224 This demo we just make a static single `etcd` server to host our Kubernetes and `etcd` master servers. 225 226 Since we are OFFLINE here most of the helping processes in CoreOS and Kubernetes are then limited. To do our setup we will then have to download and serve up our binaries for Kubernetes in our local environment. 227 228 An easy solution is to host a small web server on the DHCP/TFTP host for all our binaries to make them available to the local CoreOS PXE machines. 229 230 To get this up and running we are going to setup a simple `apache` server to serve our binaries needed to bootstrap Kubernetes. 231 232 This is on the PXE server from the previous section: 233 234 rm /etc/httpd/conf.d/welcome.conf 235 cd /var/www/html/ 236 wget -O kube-register https://github.com/kelseyhightower/kube-register/releases/download/v0.0.2/kube-register-0.0.2-linux-amd64 237 wget -O setup-network-environment https://github.com/kelseyhightower/setup-network-environment/releases/download/v1.0.0/setup-network-environment 238 wget https://storage.googleapis.com/kubernetes-release/release/v0.15.0/bin/linux/amd64/kubernetes --no-check-certificate 239 wget https://storage.googleapis.com/kubernetes-release/release/v0.15.0/bin/linux/amd64/kube-apiserver --no-check-certificate 240 wget https://storage.googleapis.com/kubernetes-release/release/v0.15.0/bin/linux/amd64/kube-controller-manager --no-check-certificate 241 wget https://storage.googleapis.com/kubernetes-release/release/v0.15.0/bin/linux/amd64/kube-scheduler --no-check-certificate 242 wget https://storage.googleapis.com/kubernetes-release/release/v0.15.0/bin/linux/amd64/kubectl --no-check-certificate 243 wget https://storage.googleapis.com/kubernetes-release/release/v0.15.0/bin/linux/amd64/kubecfg --no-check-certificate 244 wget https://storage.googleapis.com/kubernetes-release/release/v0.15.0/bin/linux/amd64/kubelet --no-check-certificate 245 wget https://storage.googleapis.com/kubernetes-release/release/v0.15.0/bin/linux/amd64/kube-proxy --no-check-certificate 246 wget -O flanneld https://storage.googleapis.com/k8s/flanneld --no-check-certificate 247 248 This sets up our binaries we need to run Kubernetes. This would need to be enhanced to download from the Internet for updates in the future. 249 250 Now for the good stuff! 251 252 ## Cloud Configs 253 254 The following config files are tailored for the OFFLINE version of a Kubernetes deployment. 255 256 These are based on the work found here: [master.yml](cloud-configs/master.yaml), [node.yml](cloud-configs/node.yaml) 257 258 To make the setup work, you need to replace a few placeholders: 259 260 - Replace `<PXE_SERVER_IP>` with your PXE server ip address (e.g. 10.20.30.242) 261 - Replace `<MASTER_SERVER_IP>` with the Kubernetes master ip address (e.g. 10.20.30.40) 262 - If you run a private docker registry, replace `rdocker.example.com` with your docker registry dns name. 263 - If you use a proxy, replace `rproxy.example.com` with your proxy server (and port) 264 - Add your own SSH public key(s) to the cloud config at the end 265 266 ### master.yml 267 268 On the PXE server make and fill in the variables `vi /var/www/html/coreos/pxe-cloud-config-master.yml`. 269 270 271 #cloud-config 272 --- 273 write_files: 274 - path: /opt/bin/waiter.sh 275 owner: root 276 content: | 277 #! /usr/bin/bash 278 until curl http://127.0.0.1:4001/v2/machines; do sleep 2; done 279 - path: /opt/bin/kubernetes-download.sh 280 owner: root 281 permissions: 0755 282 content: | 283 #! /usr/bin/bash 284 /usr/bin/wget -N -P "/opt/bin" "http://<PXE_SERVER_IP>/kubectl" 285 /usr/bin/wget -N -P "/opt/bin" "http://<PXE_SERVER_IP>/kubernetes" 286 /usr/bin/wget -N -P "/opt/bin" "http://<PXE_SERVER_IP>/kubecfg" 287 chmod +x /opt/bin/* 288 - path: /etc/profile.d/opt-path.sh 289 owner: root 290 permissions: 0755 291 content: | 292 #! /usr/bin/bash 293 PATH=$PATH/opt/bin 294 coreos: 295 units: 296 - name: 10-eno1.network 297 runtime: true 298 content: | 299 [Match] 300 Name=eno1 301 [Network] 302 DHCP=yes 303 - name: 20-nodhcp.network 304 runtime: true 305 content: | 306 [Match] 307 Name=en* 308 [Network] 309 DHCP=none 310 - name: get-kube-tools.service 311 runtime: true 312 command: start 313 content: | 314 [Service] 315 ExecStartPre=-/usr/bin/mkdir -p /opt/bin 316 ExecStart=/opt/bin/kubernetes-download.sh 317 RemainAfterExit=yes 318 Type=oneshot 319 - name: setup-network-environment.service 320 command: start 321 content: | 322 [Unit] 323 Description=Setup Network Environment 324 Documentation=https://github.com/kelseyhightower/setup-network-environment 325 Requires=network-online.target 326 After=network-online.target 327 [Service] 328 ExecStartPre=-/usr/bin/mkdir -p /opt/bin 329 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/setup-network-environment 330 ExecStartPre=/usr/bin/chmod +x /opt/bin/setup-network-environment 331 ExecStart=/opt/bin/setup-network-environment 332 RemainAfterExit=yes 333 Type=oneshot 334 - name: etcd.service 335 command: start 336 content: | 337 [Unit] 338 Description=etcd 339 Requires=setup-network-environment.service 340 After=setup-network-environment.service 341 [Service] 342 EnvironmentFile=/etc/network-environment 343 User=etcd 344 PermissionsStartOnly=true 345 ExecStart=/usr/bin/etcd \ 346 --name ${DEFAULT_IPV4} \ 347 --addr ${DEFAULT_IPV4}:4001 \ 348 --bind-addr 0.0.0.0 \ 349 --cluster-active-size 1 \ 350 --data-dir /var/lib/etcd \ 351 --http-read-timeout 86400 \ 352 --peer-addr ${DEFAULT_IPV4}:7001 \ 353 --snapshot true 354 Restart=always 355 RestartSec=10s 356 - name: fleet.socket 357 command: start 358 content: | 359 [Socket] 360 ListenStream=/var/run/fleet.sock 361 - name: fleet.service 362 command: start 363 content: | 364 [Unit] 365 Description=fleet daemon 366 Wants=etcd.service 367 After=etcd.service 368 Wants=fleet.socket 369 After=fleet.socket 370 [Service] 371 Environment="FLEET_ETCD_SERVERS=http://127.0.0.1:4001" 372 Environment="FLEET_METADATA=role=master" 373 ExecStart=/usr/bin/fleetd 374 Restart=always 375 RestartSec=10s 376 - name: etcd-waiter.service 377 command: start 378 content: | 379 [Unit] 380 Description=etcd waiter 381 Wants=network-online.target 382 Wants=etcd.service 383 After=etcd.service 384 After=network-online.target 385 Before=flannel.service 386 Before=setup-network-environment.service 387 [Service] 388 ExecStartPre=/usr/bin/chmod +x /opt/bin/waiter.sh 389 ExecStart=/usr/bin/bash /opt/bin/waiter.sh 390 RemainAfterExit=true 391 Type=oneshot 392 - name: flannel.service 393 command: start 394 content: | 395 [Unit] 396 Wants=etcd-waiter.service 397 After=etcd-waiter.service 398 Requires=etcd.service 399 After=etcd.service 400 After=network-online.target 401 Wants=network-online.target 402 Description=flannel is an etcd backed overlay network for containers 403 [Service] 404 Type=notify 405 ExecStartPre=-/usr/bin/mkdir -p /opt/bin 406 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/flanneld 407 ExecStartPre=/usr/bin/chmod +x /opt/bin/flanneld 408 ExecStartPre=-/usr/bin/etcdctl mk /coreos.com/network/config '{"Network":"10.100.0.0/16", "Backend": {"Type": "vxlan"}}' 409 ExecStart=/opt/bin/flanneld 410 - name: kube-apiserver.service 411 command: start 412 content: | 413 [Unit] 414 Description=Kubernetes API Server 415 Documentation=https://github.com/kubernetes/kubernetes 416 Requires=etcd.service 417 After=etcd.service 418 [Service] 419 ExecStartPre=-/usr/bin/mkdir -p /opt/bin 420 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/kube-apiserver 421 ExecStartPre=/usr/bin/chmod +x /opt/bin/kube-apiserver 422 ExecStart=/opt/bin/kube-apiserver \ 423 --address=0.0.0.0 \ 424 --port=8080 \ 425 --service-cluster-ip-range=10.100.0.0/16 \ 426 --etcd-servers=http://127.0.0.1:4001 \ 427 --logtostderr=true 428 Restart=always 429 RestartSec=10 430 - name: kube-controller-manager.service 431 command: start 432 content: | 433 [Unit] 434 Description=Kubernetes Controller Manager 435 Documentation=https://github.com/kubernetes/kubernetes 436 Requires=kube-apiserver.service 437 After=kube-apiserver.service 438 [Service] 439 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/kube-controller-manager 440 ExecStartPre=/usr/bin/chmod +x /opt/bin/kube-controller-manager 441 ExecStart=/opt/bin/kube-controller-manager \ 442 --master=127.0.0.1:8080 \ 443 --logtostderr=true 444 Restart=always 445 RestartSec=10 446 - name: kube-scheduler.service 447 command: start 448 content: | 449 [Unit] 450 Description=Kubernetes Scheduler 451 Documentation=https://github.com/kubernetes/kubernetes 452 Requires=kube-apiserver.service 453 After=kube-apiserver.service 454 [Service] 455 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/kube-scheduler 456 ExecStartPre=/usr/bin/chmod +x /opt/bin/kube-scheduler 457 ExecStart=/opt/bin/kube-scheduler --master=127.0.0.1:8080 458 Restart=always 459 RestartSec=10 460 - name: kube-register.service 461 command: start 462 content: | 463 [Unit] 464 Description=Kubernetes Registration Service 465 Documentation=https://github.com/kelseyhightower/kube-register 466 Requires=kube-apiserver.service 467 After=kube-apiserver.service 468 Requires=fleet.service 469 After=fleet.service 470 [Service] 471 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/kube-register 472 ExecStartPre=/usr/bin/chmod +x /opt/bin/kube-register 473 ExecStart=/opt/bin/kube-register \ 474 --metadata=role=node \ 475 --fleet-endpoint=unix:///var/run/fleet.sock \ 476 --healthz-port=10248 \ 477 --api-endpoint=http://127.0.0.1:8080 478 Restart=always 479 RestartSec=10 480 update: 481 group: stable 482 reboot-strategy: off 483 ssh_authorized_keys: 484 - ssh-rsa AAAAB3NzaC1yc2EAAAAD... 485 486 487 ### node.yml 488 489 On the PXE server make and fill in the variables `vi /var/www/html/coreos/pxe-cloud-config-slave.yml`. 490 491 #cloud-config 492 --- 493 write_files: 494 - path: /etc/default/docker 495 content: | 496 DOCKER_EXTRA_OPTS='--insecure-registry="rdocker.example.com:5000"' 497 coreos: 498 units: 499 - name: 10-eno1.network 500 runtime: true 501 content: | 502 [Match] 503 Name=eno1 504 [Network] 505 DHCP=yes 506 - name: 20-nodhcp.network 507 runtime: true 508 content: | 509 [Match] 510 Name=en* 511 [Network] 512 DHCP=none 513 - name: etcd.service 514 mask: true 515 - name: docker.service 516 drop-ins: 517 - name: 50-insecure-registry.conf 518 content: | 519 [Service] 520 Environment="HTTP_PROXY=http://rproxy.example.com:3128/" "NO_PROXY=localhost,127.0.0.0/8,rdocker.example.com" 521 - name: fleet.service 522 command: start 523 content: | 524 [Unit] 525 Description=fleet daemon 526 Wants=fleet.socket 527 After=fleet.socket 528 [Service] 529 Environment="FLEET_ETCD_SERVERS=http://<MASTER_SERVER_IP>:4001" 530 Environment="FLEET_METADATA=role=node" 531 ExecStart=/usr/bin/fleetd 532 Restart=always 533 RestartSec=10s 534 - name: flannel.service 535 command: start 536 content: | 537 [Unit] 538 After=network-online.target 539 Wants=network-online.target 540 Description=flannel is an etcd backed overlay network for containers 541 [Service] 542 Type=notify 543 ExecStartPre=-/usr/bin/mkdir -p /opt/bin 544 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/flanneld 545 ExecStartPre=/usr/bin/chmod +x /opt/bin/flanneld 546 ExecStart=/opt/bin/flanneld -etcd-endpoints http://<MASTER_SERVER_IP>:4001 547 - name: docker.service 548 command: start 549 content: | 550 [Unit] 551 After=flannel.service 552 Wants=flannel.service 553 Description=Docker Application Container Engine 554 Documentation=http://docs.docker.io 555 [Service] 556 EnvironmentFile=-/etc/default/docker 557 EnvironmentFile=/run/flannel/subnet.env 558 ExecStartPre=/bin/mount --make-rprivate / 559 ExecStart=/usr/bin/docker -d --bip=${FLANNEL_SUBNET} --mtu=${FLANNEL_MTU} -s=overlay -H fd:// ${DOCKER_EXTRA_OPTS} 560 [Install] 561 WantedBy=multi-user.target 562 - name: setup-network-environment.service 563 command: start 564 content: | 565 [Unit] 566 Description=Setup Network Environment 567 Documentation=https://github.com/kelseyhightower/setup-network-environment 568 Requires=network-online.target 569 After=network-online.target 570 [Service] 571 ExecStartPre=-/usr/bin/mkdir -p /opt/bin 572 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/setup-network-environment 573 ExecStartPre=/usr/bin/chmod +x /opt/bin/setup-network-environment 574 ExecStart=/opt/bin/setup-network-environment 575 RemainAfterExit=yes 576 Type=oneshot 577 - name: kube-proxy.service 578 command: start 579 content: | 580 [Unit] 581 Description=Kubernetes Proxy 582 Documentation=https://github.com/kubernetes/kubernetes 583 Requires=setup-network-environment.service 584 After=setup-network-environment.service 585 [Service] 586 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/kube-proxy 587 ExecStartPre=/usr/bin/chmod +x /opt/bin/kube-proxy 588 ExecStart=/opt/bin/kube-proxy \ 589 --etcd-servers=http://<MASTER_SERVER_IP>:4001 \ 590 --logtostderr=true 591 Restart=always 592 RestartSec=10 593 - name: kube-kubelet.service 594 command: start 595 content: | 596 [Unit] 597 Description=Kubernetes Kubelet 598 Documentation=https://github.com/kubernetes/kubernetes 599 Requires=setup-network-environment.service 600 After=setup-network-environment.service 601 [Service] 602 EnvironmentFile=/etc/network-environment 603 ExecStartPre=/usr/bin/wget -N -P /opt/bin http://<PXE_SERVER_IP>/kubelet 604 ExecStartPre=/usr/bin/chmod +x /opt/bin/kubelet 605 ExecStart=/opt/bin/kubelet \ 606 --address=0.0.0.0 \ 607 --port=10250 \ 608 --hostname-override=${DEFAULT_IPV4} \ 609 --api-servers=<MASTER_SERVER_IP>:8080 \ 610 --healthz-bind-address=0.0.0.0 \ 611 --healthz-port=10248 \ 612 --logtostderr=true 613 Restart=always 614 RestartSec=10 615 update: 616 group: stable 617 reboot-strategy: off 618 ssh_authorized_keys: 619 - ssh-rsa AAAAB3NzaC1yc2EAAAAD... 620 621 622 ## New pxelinux.cfg file 623 624 Create a pxelinux target file for a _slave_ node: `vi /tftpboot/pxelinux.cfg/coreos-node-slave` 625 626 default coreos 627 prompt 1 628 timeout 15 629 630 display boot.msg 631 632 label coreos 633 menu default 634 kernel images/coreos/coreos_production_pxe.vmlinuz 635 append initrd=images/coreos/coreos_production_pxe_image.cpio.gz cloud-config-url=http://<pxe-host-ip>/coreos/pxe-cloud-config-slave.yml console=tty0 console=ttyS0 coreos.autologin=tty1 coreos.autologin=ttyS0 636 637 And one for the _master_ node: `vi /tftpboot/pxelinux.cfg/coreos-node-master` 638 639 default coreos 640 prompt 1 641 timeout 15 642 643 display boot.msg 644 645 label coreos 646 menu default 647 kernel images/coreos/coreos_production_pxe.vmlinuz 648 append initrd=images/coreos/coreos_production_pxe_image.cpio.gz cloud-config-url=http://<pxe-host-ip>/coreos/pxe-cloud-config-master.yml console=tty0 console=ttyS0 coreos.autologin=tty1 coreos.autologin=ttyS0 649 650 ## Specify the pxelinux targets 651 652 Now that we have our new targets setup for master and slave we want to configure the specific hosts to those targets. We will do this by using the pxelinux mechanism of setting a specific MAC addresses to a specific pxelinux.cfg file. 653 654 Refer to the MAC address table in the beginning of this guide. Documentation for more details can be found [here](http://www.syslinux.org/wiki/index.php/PXELINUX). 655 656 cd /tftpboot/pxelinux.cfg 657 ln -s coreos-node-master 01-d0-00-67-13-0d-00 658 ln -s coreos-node-slave 01-d0-00-67-13-0d-01 659 ln -s coreos-node-slave 01-d0-00-67-13-0d-02 660 661 662 Reboot these servers to get the images PXEd and ready for running containers! 663 664 ## Creating test pod 665 666 Now that the CoreOS with Kubernetes installed is up and running lets spin up some Kubernetes pods to demonstrate the system. 667 668 See [a simple nginx example](../../../docs/user-guide/simple-nginx.md) to try out your new cluster. 669 670 For more complete applications, please look in the [examples directory](../../../examples/). 671 672 ## Helping commands for debugging 673 674 List all keys in etcd: 675 676 etcdctl ls --recursive 677 678 List fleet machines 679 680 fleetctl list-machines 681 682 Check system status of services on master: 683 684 systemctl status kube-apiserver 685 systemctl status kube-controller-manager 686 systemctl status kube-scheduler 687 systemctl status kube-register 688 689 Check system status of services on a node: 690 691 systemctl status kube-kubelet 692 systemctl status docker.service 693 694 List Kubernetes 695 696 kubectl get pods 697 kubectl get nodes 698 699 700 Kill all pods: 701 702 for i in `kubectl get pods | awk '{print $1}'`; do kubectl stop pod $i; done 703 704 705 <!-- BEGIN MUNGE: GENERATED_ANALYTICS --> 706 []() 707 <!-- END MUNGE: GENERATED_ANALYTICS -->