Just one question: do yoi have any drawings for Tsuru Architecture?

For example, will be possible to split each components on multiple servers, like one git server, 5 node servers (with docker), 6 mysql servers, and so on?

AFAIK , the standard installation will place everything on the same server with capability to use external mysql servers. But what If I would like to have redundancy and multiple nodes?

I'm using docker provisioner and getting this error with debug set. The images are in the docker registry.

create container for app abyss, based on image xxxxx.domain.com:8080/tsuru/abyss

error on creating container in docker abyss - No such image

Here is docker image list

REPOSITORY                                    TAG
 IMAGE ID            CREATED             VIRTUAL SIZE
xxxxx.domain.com:8080/tsuru/abyss    latest
 515a19752542        10 minutes ago      634 MB

Here is the last of tsuru config

registry: xxxxx.domain.com:8080
repository-namespace: tsuru

Docker registry files

root@p:/var/lib/docker-registry/repositories/tsuru/abyss/tags# ls -la
total 12
drwxr-xr-x 2 root root 4096 Mar 22 04:23 .
drwxr-xr-x 3 root root 4096 Mar 22 04:23 ..
-rw-r--r-- 1 root root   66 Mar 22 04:23 latest

Currently, we use post-receive to replicate code across units of the application, we should use pre-receive. This is a large change, but for good. By using pre-receive, Tsuru would reject any push that won't deploy, preserving the state of the bare repository.

In order to make this change, cloneRepository handler would need to receive the object ID (the commit identifier). The pre-receive hook would look something like this:

#!/bin/bash -el
app_dir=${PWD##*/}
app_name=${app_dir/.git/}
url="${TSURU_HOST}/apps/${app_name}/repository/clone"
hash=`read line | awk '{print $2}'`
curl -H "Authorization: bearer ${TSURU_TOKEN}" -d "version=${hash}" -s -N --max-time 1800 $url

But then we need to decide how provisioners will handle that. Currently, the provisioner is responsible for deployment. We can pass another parameter to the Deploy method, "version":

type Provisioner interface {
    // other methods

   // Deploy updates the code of the app in units to match the given version, logging progress in the given writer.
    Deploy(app provision.App, version string, w io.Writer) error
}

And the provisioner decides what "version" means and how to handle it. Currently, it would be a git commit identifier, but the name support other things, like a SVN or mercurial commit, or just a binary version. Again, it's up to the provisioner, tsuru does not care.

This change would automatically fix #442.

I followed this howto: http://docs.tsuru.io/en/latest/services/mysql-example.html

When I want to add instance, I get error: http://pastebin.com/raw.php?i=BB11hZdC

I tried:

root@a:~# curl -d 'name=blog' http://mysql-api.example.com:8888/resources
int() argument must be a string or a number, not 'NoneType'root@a:~#

It's very similar to this issue: https://github.com/globocom/tsuru/issues/667, but I can't find there solution.

Thanks for help

I'm testing PHP platform. Every time that I build a new image and I push it to the central registry (we are using a custom registry), the push will be ok but I'm unable to create a new tsuru app from that image.

The old image is always used, and I have to create every time a brand new platform. Do you have any image cache somewhere? Any way to force a full rebuild for the platform ignoring cache or something similar ?

I think that tsusu see a previously known platform name and use the old image instead downloading the new image from the repository

On Google Cloud, these are the latest lines when running tsuru install

Tsuru API successfully installed!
Bootstrapping Tsuru API...adding target
New target cloudster-test -> http://1.2.3.4:8080 added to target list and defined as the current target
log in with default user: [email protected]: 
Successfully logged in!
adding pool
Pool successfully registered.
creating node 1/3...
Error: Error bootstrapping tsuru: failed to create node: 404 page not found

I'm following this guide: http://docs.tsuru.io/en/latest/services/mysql-example.html

bot something is not working properly. My app is running, but is not responding to HTTP request

I think that gunicorn is not running properly. How can I debug gunicorn output?

I've reinstalled a docker node and created a brand new tsuru/php container Now, when pushing with git, i received this error:

remote: No such container: 0cbe3eba580b

I think that samewhere is stored the old image ID. How can I remove this "cache" ?

Here's my config for docker: block:

    sharedfs:
        hostdir: /mnt/storage/dockerstor
        mountpoint: /storage
        app-isolation: true

When I try to re-start tsuru-server-api, it crashed without any output.

Additional information:

Docker version 1.3.1, Tsuru-server: latest On Ubuntu 14.04 x64

    $ ls -la /mnt/storage/dockerstor
    total 0
    drwxrwxrwx 1 tsuru uulm 0 Nov 25 15:38 .
    drwxrwxr-x 1 root  root 0 Nov 24 14:17 ..

And in fstab :

192.168.26.190:6789:/   /mnt/storage    ceph    name=foo,secret=secrt_string        0  2

I met Tsuru few days ago, I could have tried to install before and works, but now after the installation of Tsuru-server when I try to create some command by Tsuru-client get this error:

Mac-Pro-de-jadson:teste jadson$ tsuru login [email protected]
Password: 
Error: <html>
  <head>
    <title>No Application Configured</title>
    <style>
      body { background-color: #fff; margin: 0px; padding-top: 280px; text-align: center; }
      h1   { font-family: 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 24px; font-weight: normal; }
      h2   { font-family: 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; color: #999; font-weight: normal; }
    </style>
  </head>
  <body>
    <div class="message">
      <h1>No Application Configured</h1>
      <h2>This domain is not associated with an application.</h2>
    </div>
  </body>
</html>

is the same page that is on port 80 when the access IP server directly by the browser.

Does anyone have any idea how to solve?

Leveraging the question: Dashboard Tsuru is already present in the normal installation, or is it necessary to create an "app" within the tsuru, with the client or admin?

PS: I'm following the installation guide that has on the tsuru.io site, using ubuntu 12.04 (but the error also happens with ubuntu 13.10) - I did this in the DigitalOcean and Vagrant, both the same error.

Thanks!

The docs at http://docs.tsuru.io/en/latest/docker.html refer to /vagrant in the Gandalf and API server section as well as to an unknown shell variable host_ip.

Bumps github.com/aws/aws-sdk-go from 1.28.2 to 1.33.0.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Tsuru API works well with CLIs and backends for frontends, each oauth2 token must be stored in database and oauth2 provider validate these tokens, but we need to think to do a step further in terms of usage, actually we can't negotiate a token with auth provider using implicit flow and pass through Tsuru API server, the server must know the token previously.

To achieve a support to use API using a Single Page Applications, let's introduce JWK support.

JWK is a standard to validate tokens without calling a identity provider, https://www.rfc-editor.org/rfc/rfc7517.

what will happen:

The browser will negotiate a JWT token with an identity provider, the token will pass as a Authorization header, tsuru will validate the token based on standard claims, and response directly to the user browser.

Strong Important Validations:

• use sub claim as a user email
• validate expiration of token
• validate audience field
• validate by public key of JWK

-> Vulnerability found in sphinx version 1.7.4 Vulnerability ID: 38330 Affected spec: <3.0.4 ADVISORY: Sphinx 3.0.4 updates jQuery version from 3.4.1 to 3.5.1 for security reasons. CVE-2020-11022 For more information, please visit https://pyup.io/vulnerabilities/CVE-2020-11022/38330/

-> Vulnerability found in sphinx version 1.7.4 Vulnerability ID: 45775 Affected spec: <3.0.4 ADVISORY: Sphinx 3.0.4 updates jQuery version from 3.4.1 to 3.5.1 for security reasons. CVE-2020-11023 For more information, please visit https://pyup.io/vulnerabilities/CVE-2020-11023/45775/

I have documented the steps required for installing Tsuru using make local in linux. I have tried to account for small obstacles and workarounds encountered while running the makefile.

It partially fixes #2575, as macOS users may face some different errors while documenting, hope it helps!

Close #2576

Enable alias on register new command.

When register a new command, now you can inform alias field to generate a new alias.

e.g. tsuru app log and tsuru app logs call the same command run.

Is your feature request related to a problem? Please describe. N/A

Describe the solution you'd like

Be able to run tsuru cli commands with a alias eg. tsuru log same with tsuru logs

Describe alternatives you've considered.

Create a alias on bash.

Additional context

I'm used to use logs as command for log view, like in docker logs. It would be awesome being able to call with log or logs