I'm trying to build an Amazon AMI centos using Packer. I am using the amazon-chroot builder.
The ami exists, but I am getting the build error
[root#ip-10-32-11-16 retel-base]# packer build retel-base.json
amazon-chroot output will be in this color.
==> amazon-chroot: Gathering information about this EC2 instance...
==> amazon-chroot: Inspecting the source AMI...
==> amazon-chroot: Couldn't find root device!
Build 'amazon-chroot' errored: Couldn't find root device!
==> Some builds didn't complete successfully and had errors:
--> amazon-chroot: Couldn't find root device!
==> Builds finished but no artifacts were created.
cat retel-base.json
{
"variables": {
"ACCESS_KEY_ID": "{{env `ACCESS_KEY_ID`}}",
"SECRET_ACCESS_KEY": "{{env `SECRET_ACCESS_KEY`}}"
},
"builders": [{
"type": "amazon-chroot",
"access_key": "{{user `ACCESS_KEY_ID`}}",
"secret_key": "{{user `SECRET_ACCESS_KEY`}}",
"source_ami":"ami-a40df4cc",
"ami_name": "base image built with packer {{timestamp}}"
}]
}
I think this might be to do with a mismatch between the name of the root device and the block device mapping.
In the official CentOS AMIs, the root device is named /dev/sda but the block device mapping only lists /dev/sda1, which is apparently a partition on the root device.
The Aminator by Netflix has a similar problem with partitioned volumes: https://github.com/Netflix/aminator/issues/129
Related
I have installed Docker Desktop for windows Docker version 18.09.2, build 6247962, and I'm not able to build and image. Even a docker search does not seem to work.
The error message (for example, when doing a docker search) is:
Error response from daemon: Get https://index.docker.io/v1/search?q=ubuntu&n=25: proxyconnect tcp: dial tcp 172.17.14.133:3128: connect: no route to host
My office is behind a proxy. So on the "Proxies" settings of DockerDesktop I set http://172.17.14.133:3128 for both HTTP and HTTTPS. But it still does not seem to work.
I have defined some ENV variables (both user and system) like this:
HTTPS_PROXY=http://proxypmi.tradyso.com:3128
HTTP_PROXY=http://proxypmi.tradyso.com:3128
Also:
C:\Users\my.user\AppData\Roaming\Docker\http_proxy.json:
{
"http": "http://172.17.14.133:3128",
"https": "http://172.17.14.133:3128",
"exclude": null,
"transparent_http_ports": [],
"transparent_https_ports": []
}
C:\Users\my.user\AppData\Roaming\Docker\settings.json:
{
"settingsVersion": 1,
"autoStart": false,
"checkForUpdates": true,
"analyticsEnabled": false,
"displayedWelcomeWhale": true,
"displayed14393Deprecation": false,
"displayRestartDialog": true,
"displaySwitchWinLinContainers": true,
"latestBannerKey": "",
"debug": false,
"memoryMiB": 2048,
"swapMiB": 1024,
"cpus": 2,
"diskPath": null,
"diskSizeMiB": 64000000000,
"networkCIDR": "10.0.75.0/24",
"proxyHttpMode": true,
"overrideProxyHttp": "http://172.17.14.133:3128",
"overrideProxyHttps": "http://172.17.14.133:3128",
"overrideProxyExclude": null,
"useDnsForwarder": true,
"dns": "10.44.24.10",
"kubernetesEnabled": false,
"showKubernetesSystemContainers": false,
"kubernetesInitialInstallPerformed": false,
"cliConfigCreationDate": "03/22/2019 12:23:58",
"linuxDaemonConfigCreationDate": "03/22/2019 12:22:19",
"windowsDaemonConfigCreationDate": null,
"versionPack": "default",
"sharedDrives": {},
"executableDate": "",
"useWindowsContainers": false,
"swarmFederationExplicitlyLoggedOut": false,
"activeOrganizationName": null,
"exposeDockerAPIOnTCP2375": false
}
C:\Users\my.user\.docker\config.json:
{
"stackOrchestrator": "swarm",
"auths": {},
"credsStore": "wincred",
"proxies":
{
"default":
{
"httpProxy": "http://172.17.14.133:3128",
"httpsProxy": "http://172.17.14.133:3128",
"noProxy": ""
}
}
}
I also tried passing build-arg to tocker build:
docker build --build-arg HTTP_PROXY=http://172.17.14.133:3128 --build-arg HTTPS_PROXY=http://172.17.14.133:3128 ...
Finally, on the Docker Desktop network configuration, I have tried with DNSs both "Automatic" and Manual (Using my corporate dns servers)
None of this has worked.
Any hint on what should I do?
Thank you.
A collegue found out the problem:
By default, the bridge network that docker creates uses the same subnet as our office (172.17.0.0/16) and that causes trouble with the proxy ip address (172.17.14.133).
To solve this:
[Edit]: for a simpler method, use the following:
On daemon configuration, add "bip": "new_subbet". For example: "bip": "172.19.0.1/16". Then, restart docker.
Now, you won't even need to pass the extra --network="docker_gwbridge" parameter to the commands.
This shorter solution may not work with older versions of Docker for windows, so you may consider the original answer if this option does not work.
[Edit]: original method for old versions of docker for windows:
The bridge network cannot be deleted, but We can tell docker not to create it.
Go to Daemon Settings, Advanced => add "bridge": "none", to the configuration
After applying changes, Docker will restart and now We will be able to create our own custom bridge network
In this example, We are going to use (172.19.0.0/16)
Open a console and write:
docker network create --subnet=172.19.0.0/16 --gateway 172.19.0.1 -o com.docker.network.bridge.enable_icc=true -o com.docker.network.bridge.name=docker_gwbridge -o com.docker.network.bridge.enable_ip_masquerade=true docker_gwbridge
Then we can do a docker ls for checking that the previous command was successful:
$ docker network ls
NETWORK ID NAME DRIVER SCOPE
2a3635a49ffa docker_gwbridge bridge local
4e9ec758ee9f host host local
737c6c5fc82b none null local
Now do a docker search ubuntu (for example). It should be able to connect to the internet and fetch the images.
Important: From now on, some commands that need internet access will need to specify the new docker network with the extra parameter --network="docker_gwbridge"
For example a docker build command could be:
docker build --network="docker_gwbridge" --build-arg http_proxy=http://172.17.14.133:3128 --build-arg https_proxy=http://172.17.14.133:3128 -t foobar .
I have a K8 cluster on GCP running elasticsearch. Now I need to create a backup.
I've installed the GCS-plugin on my pods in stateful-set and tried setting it up with the following documentation:
https://github.com/elastic/elasticsearch/blob/master/docs/plugins/repository-gcs.asciidoc
When I try to configure a repository to use credentials stored in keystore I get the following response back:
{
"error": {
"root_cause": [
{
"type": "repository_exception",
"reason": "[my_backup] repository type [gcs] does not exist"
}
],
"type": "repository_exception",
"reason": "[my_backup] repository type [gcs] does not exist"
},
"status": 500
}
Any lead would be helpful, thanks!
I think the problem is that I can't install the plugin on the nodes, so I’ve installed it on the pods instead. And that the installation is not persistent after I restart the pods. So to make the installation persist on K8 I needed to build a custom image that installs the plugin. A bit tricky, but the plugin seems to be intended for GCE. So I decided to move from K8 to a managed instance group on GCE instead.
All,
I'm learning Docker on my Windows Desktop 10. Windows is Pro edition and Docker is 18.09.
When I run the below -
docker run -it nanoserver/iis -v
C:\ProgramData\Docker\volumes\vol01:C:\vol01 cmd.exe
I get the below error -
docker: Error response from daemon: container
5a1229eca277cbddeefd5637e69554458003c54be3f30cc44ca41c8fa68a4a94
encountered an error during CreateProcess: failure in a Windows system
call: The system cannot find the file specified. (0x2) [Event Detail:
Provider: 00000000-0000-0000-0000-000000000000] extra info:
{"CommandLine":"-v C:\ProgramData\Docker\volumes\vol01:C:\vol01
cmd.exe","WorkingDirectory":"C:\","EmulateConsole":true,"CreateStdInPipe":true,"CreateStdOutPipe":true,"ConsoleSize":[63,237]}.
The volume does exist -
docker volume inspect vol01
[
{
"CreatedAt": "2018-12-26T03:01:01-05:00",
"Driver": "local",
"Labels": {},
"Mountpoint": "C:\ProgramData\Docker\volumes\vol01\_data",
"Name": "vol01",
"Options": {},
"Scope": "local"
} ]
I don't know what is wrong. Can someone point me in the right direction?
Thanks,
rgn
You should declare volumes before image name for docker run. Consider:
docker run -it -v C:\ProgramData\Docker\volumes\vol01:C:\vol01 nanoserver/iis cmd.exe
I am attempting to build a vagrant .box using the qemu/kvm hypervisor using packer. I am using the vagrant plugin vagrant-libvirt.
In my packer base.json file, I am using the vagrant post-processor, like
"post-processors": [
{
"type": "vagrant",
"output": "output/my-ubuntu_14.04_packer.box"
}
],
When I build the box however, I get this error in the logs:
2015/12/22 20:11:57 ui error: Build 'qemu' errored: 1 error(s) occurred:
* Post-processor failed: Unknown artifact type, can't build box: transcend.qemu
I've looked in to this issue, and it does seem to be caused by that post-processor configuration. For example, according to the conversation here, qemu did not support the vagrant post processor at that time. However, looking at this it should work now that this pull request is merged.
In fact, the documentation for the vagrant post-processor seems lists qemu as a supported provider.
I believe this is because I am currently using Packer v0.6.0 (released May 2, 2014). The vagrant post-processor wasn't supported until Packer v0.7.2 (released Oct 28, 2014), so an upgrade would be necessary to achieve this.
I'm trying to mup deploy the todos example of Meteor to a Vagrant VM running Ubuntu 14.04 LTS x64.
Meteor Up supports Windows (I'm on Windows 7):
You can use install and use Meteor Up from Linux, Mac and Windows.
This is my c:\code\todos\mup.json:
{
"servers": [
{
"host": "127.0.0.1",
"port": 2222,
"username": "vagrant",
"password": "vagrant"
}
],
"setupMongo": true,
"setupNode": true,
"nodeVersion": "0.12.4",
"setupPhantom": true,
"enableUploadProgressBar": false,
"appName": "todos-app",
"app": "/code/todos",
"env": {
"ROOT_URL": "http://127.0.0.1",
"PORT": "3001", // The port you want to bind to on your server.
"UPSTART_UID": "vagrant" // The user you want to run meteor as.
},
"deployCheckWaitTime": 30
}
My Vagrant VM is up and PuTTYTray is connected via vagrant:vagrant#127.0.0.7:2222. Yet mup deploy fails:
C:\code\todos>mup deploy
Meteor Up: Production Quality Meteor Deployments
------------------------------------------------
" Checkout Kadira!
It's the best way to monitor performance of your app.
Visit: https://kadira.io/mup "
Building Started: /code/todos
? Can't build for mobile on Windows. Skipping the following platforms:
android, ios
Started TaskList: Deploy app 'todos-app' (linux)
[127.0.0.1] - Uploading bundle
events.js:85
throw er; // Unhandled 'error' event
^
Error: connect ECONNREFUSED
at exports._errnoException (util.js:746:11)
at TCPConnectWrap.afterConnect [as oncomplete] (net.js:1010:19)
Same for mup setup. And in the VM mup deploy encounters a "weird error".
Should I downgrade mup?
Downgrading mup doesn't help:
C:\code\todos>npm install -g mup#0.9.7
C:\Users\Cees.Timmerman\AppData\Roaming\npm\mup -> C:\Users\Cees.Timmerman\AppData\Roaming\npm\node_modules\mup\bin\mup
mup#0.9.7 C:\Users\Cees.Timmerman\AppData\Roaming\npm\node_modules\mup
├── colors#0.6.2
├── underscore#1.7.0
├── uuid#1.4.2
├── async#0.9.2
├── rimraf#2.4.0 (glob#4.5.3)
├── cjson#0.3.1 (jsonlint#1.6.0)
└── nodemiral#0.3.11 (debug#0.7.4, ejs#0.8.8, handlebars#1.0.12)
C:\code\todos>mup -v
Meteor Up: Production Quality Meteor Deployments
------------------------------------------------
sshpass required for password based authentication: refer http://git.io/_vHbvQ
C:\code\todos>mup deploy
Meteor Up: Production Quality Meteor Deployments
------------------------------------------------
sshpass required for password based authentication: refer http://git.io/_vHbvQ
SSHPass doesn't work on Windows.
Manually setting up my Meteor app worked with help from this answer.