gitlab Rack_Attack settings - gitlab

I have a gitlab CE installation with Rack_Attack enabled, but I am not able to find the current values set for the following parameters - max_retry, findtime, and bantime.
So, where and how can I check the values of these parameters in my gitlab installation?

If you installed the Omnibus GitLab version, the settings can be found in /etc/gitlab/gitlab.rb:
gitlab_rails['rack_attack_git_basic_auth'] = {
'enabled' => true,
'ip_whitelist' => ["127.0.0.1"],
'maxretry' => 10,
'findtime' => 60,
'bantime' => 3600
}
If you installed GitLab from source, the settings can be found in config/initializers/rack_attack.rb.
From the Documentation.

Related

Creating master and slaves DNS with puppet module camptocamp bind

Trying to create a master and slave (redundancy) DNS with puppet module camptocamp bind. In slave profile, I've set transfer_source => '192.168.1.20' to masters ip: 192.168.1.20. It should then synchronize and copy dns records from master to the slave.
But I got complaints about that it could only be set to slave zones. I've followed the README from puppet forge for the module: https://forge.puppet.com/camptocamp/bind/readme
dnsmaster.pp
class profile::dnsbind::server {
include 'bind'
bind::zone {'example.com':
ensure => 'present',
zone_contact => 'contact.example.com',
zone_ns => ['ns0.example.com'],
zone_serial => '2012112901',
zone_ttl => '604800',
zone_origin => 'example.com',
}
bind::a { 'example.com':
ensure => 'present',
zone => 'example.com',
ptr => false,
hash_data => {
'host1' => { owner => '192.168.0.1', },
'host2' => { owner => '192.168.0.2', },
},
}
}
dnsslave.pp
class profile::dnsbind::server_slave {
include 'bind'
bind::zone {'example.com':
ensure => 'present',
zone_contact => 'contact.example.com',
zone_ns => ['ns0.example.com'],
zone_serial => '2012112901',
zone_ttl => '604800',
zone_origin => 'example.com',
transfer_source => '192.168.1.20',
}
}
The error message:
Error: Could not retrieve catalog from remote server: Error 500 on SERVER: Server Error: Evaluation Error: Error while evaluating a Resource Statement, Evaluation Error: Error while evaluating a Function Call, Zone 'example.com': transfer_source can be set only for slave zones! at /etc/puppetlabs/code/environments/production/modules/bind/manifests/zone.pp:80:5 at /etc/puppetlabs/code/environments/production/manifests/profile_dns2.pp:5 on node centos7-3
Warning: Not using cache on failed catalog
Error: Could not retrieve catalog; skipping run
It should then synchronize and copy dns records from master to the
slave.
But I got complaints about that it could only be set to slave zones.
Evidently, the module does not recognize that you're trying to configure a slave zone. How do you suppose it would know? Well, apparently not from the presence of a transfer_source property.
I've followed the README from puppet forge for the module:
https://forge.puppet.com/camptocamp/bind/readme
I'll believe that you started by pulling the example zone definition (for a master zone) from the readme, and I grant you that this module's docs are kinda shoddy. But do nevertheless consider actually reading the docs thoroughly, not just skimming them. If you had done, you would have found documentation for the zone_type parameter immediately following the the documentation for the transfer_source parameter:
$zone_type = master
Specify if the zone is master/slave/forward.
Use this to specify that you're configuring a slave zone.

Chef: Modify existing resource from another cookbook

I have two cookbooks: elasticsearch and curator.
Elasticsearch cookbook installs and configure an elasticsearch. The following resource (from elasticsearch cookbook), has to be modified from curator cookbook:
elasticsearch_configure 'elasticsearch' do
configuration ({
'http.port' => port,
'cluster.name' => cluster_name,
'node.name' => node_name,
'bootstrap.memory_lock' => false,
'discovery.zen.minimum_master_nodes' => 1,
'xpack.monitoring.enabled' => true,
'xpack.graph.enabled' => false,
'xpack.watcher.enabled' => true
})
end
I need to modify it on curator cookbook and add a single line:
'path.repo' => (["/backups/s3_currently_dev", "/backups/s3_currently", "/backups/s3_daily", "/backups/s3_weekly", "/backups/s3_monthly"])
How I can do that?
I initially was going to point you to the chef-rewind gem, but that actually points to the edit_resource provider that is now built into Chef. A basic example of this:
# cookbook_a/recipes/default.rb
file 'example.txt' do
content 'this is the initial content'
end
.
# cookbook_b/recipes/default.rb
edit_resource! :file, 'example.txt' do
content 'modified content!'
end
If both of these are in the Chef run_list, the actual content within example.txt is that of the edited resource, modified content!.
Without fully testing your case, I'm assuming the provider can be utilized the same way, like so:
edit_resource! :elasticsearch_configure, 'elasticsearch' do
configuration ({
'http.port' => port,
'cluster.name' => cluster_name,
'node.name' => node_name,
'bootstrap.memory_lock' => false,
'discovery.zen.minimum_master_nodes' => 1,
'xpack.monitoring.enabled' => true,
'xpack.graph.enabled' => false,
'xpack.watcher.enabled' => true,
'path.repo' => ["/backups/s3_currently_dev", "/backups/s3_currently", "/backups/s3_daily", "/backups/s3_weekly", "/backups/s3_monthly"]
})
end

Puppet can't find class firewall

I have a basic puppet install using this tutorial https://www.digitalocean.com/community/tutorials/how-to-install-puppet-4-on-ubuntu-16-04
When I run /opt/puppetlabs/bin/puppet agent --test on my node I get
Error: Could not retrieve catalog from remote server: Error 500 on SERVER: Server Error: Error while evaluating a Resource Statement. Could not find declared class firewall at /etc/puppetlabs/code/environments/production/manifests/site.pp:7:1 on node mark-inspiron.
On my node:
/opt/puppetlabs/bin/puppet module list
returns
/etc/puppetlabs/code/environment/production/modules
----- puppetlabs-firewall (v1.9.0)
On my puppet master at /etc/puppetlabs/code/environments/production/manifests/site.pp:
file {'/tmp/it_works.txt': # resource type file and filename
ensure => present, # make sure it exists
mode => '0644', # file permissions
content => "It works on ${ipaddress_eth0}!\n", # Print the eth0 IP fact
}
class { 'firewall': }
resources { 'firewall':
purge => true,
}
firewall { "051 asterisk-set-rate-limit-register":
string => "REGISTER sip:",
string_algo => "bm",
dport => '5060',
proto => 'udp',
recent => 'set',
rname => 'VOIPREGISTER',
rsource => 'true';
}
firewall { "052 asterisk-drop-rate-limit-register":
string => "REGISTER sip:",
string_algo => "bm",
dport => '5060',
proto => 'udp',
action => 'drop',
recent => 'update',
rseconds => '600',
rhitcount => '5',
rname => 'VOIPREGISTER',
rsource => true,
rttl => true;
}
The file part works but not firewall.
You need to install the modules on your master in a master setup with Puppet. They need to be somewhere in your modulepath. You can either place it in the modules directory within your $codedir (normally /etc/puppetlabs/code/modules) or in your directory environment modules directory (likely /etc/puppetlabs/code/environments/production/modules in your case since your cited site.pp is there). If you defined additional module paths in your environment.conf, then you can also place the modules there.
You can install/deploy them with a variety of methods, such as librarian-puppet, r10k, or code-manager (in Enterprise). However, the easiest method for you would be puppet module install puppetlabs-firewall on the master. Your Puppet catalog will then find the firewall class during compilation.
On a side note, that:
resources { 'firewall':
purge => true,
}
will remove any changes to associated firewall configurations (as defined by Puppet's knowledge of the system firewall configuration according to the module's definition of what the resource manages) that are not managed by Puppet. This is nice for eliminating local changes that people make, but it can also have interesting side effects, so be careful.

Gitlab SAML Configuration - 404 on metadata

Question regarding SAML configuration.
I'm currently running Gitlab 9.1 CE edition on CentOs 7. I have an Apache instance on the front end for a reverse proxy to Gitlab handling http(s)
My gitlab.rb has the following configured
external_url 'http://external.apache.server/gitlab/'
gitlab_rails['omniauth_enabled'] = true
gitlab_rails['omniauth_allow_single_sign_on'] = ['saml']
gitlab_rails['omniauth_auto_sign_in_with_provider'] = 'saml'
gitlab_rails['omniauth_block_auto_created_users'] = false
# gitlab_rails['omniauth_auto_link_ldap_user'] = false
gitlab_rails['omniauth_auto_link_saml_user'] = true
# gitlab_rails['omniauth_external_providers'] = ['twitter', 'google_oauth2']
# gitlab_rails['omniauth_providers'] = [
# {
# "name" => "google_oauth2",
# "app_id" => "YOUR APP ID",
# "app_secret" => "YOUR APP SECRET",
# "args" => { "access_type" => "offline", "approval_prompt" => "" }
# }
# ]
In order to setup SAML my provider is asking for the information returned from http://external.apache.server/gitlab/users/auth/saml/metadata which returns a 404.
In reading the SAML documentation, it mentions that Gitlab needs to be configured for SSL, not sure if this is why the URL mentioned above is returning a 404.
The problem with enabling SSL is that my external URL is already providing that and if I use it as is https://external.apache.server then Gitlab is looking for key/cert for that domain on the box which doesn't seem correct. I don't want to change the external URL as it should be fronted by Apache. Bit confused on what the proper configuration should be.
Thanks

Containing a defined resource type in Puppet

I'm trying to create an instance of a defined resource type (::apt::ppa) that comes before other resources. I am using the PuppetLabs Apt Module.
When adding a new repository via the module, the defined type contains an exec statement that notifies apt::update so that any packages that might be required can be installed correctly. However, when I run my below code, the notify gets scheduled after I attempt to install Java, thereby causing the Java install to fail. I've tried putting anchors around the apt::ppa declaration, but that doesn't help. What else can I do?
class rap::java(
$version = '7',
) {
$package = "oracle-java${version}-installer"
apt::ppa { 'ppa:webupd8team/java': } ->
exec { 'accept-java-license':
command => "/bin/echo ${package} shared/accepted-oracle-license-v1-1 select true | /usr/bin/sudo /usr/bin/debconf-set-selections",
unless => "/usr/bin/debconf-show ${package} | grep 'shared/accepted-oracle-license-v1-1: true'",
} ->
class { '::java':
package => $package,
distribution => 'oracle-jre',
}
file_line { 'java_environment':
path => '/etc/environment',
line => "JAVA_HOME=\"/usr/lib/jvm/java-${version}-oracle\"",
}
}
I believe the issue is that you need to include the apt class within the class you've made to get the ordering right.
This works for me on a new Precise box:
class rap::java(
$version = '7',
) {
$package = "oracle-java${version}-installer"
include apt
apt::ppa { 'ppa:webupd8team/java':
package_manage => true,
}
exec { 'accept-java-license':
command => "/bin/echo ${package} shared/accepted-oracle-license-v1-1 select true | /usr/bin/sudo /usr/bin/debconf-set-selections",
unless => "/usr/bin/debconf-show ${package} | grep 'shared/accepted-oracle-license-v1-1: true'",
}
class { '::java':
package => $package,
distribution => 'oracle-jre',
require => [
Apt::Ppa['ppa:webupd8team/java'],
Exec["accept-java-license"],
]
}
file_line { 'java_environment':
path => '/etc/environment',
line => "JAVA_HOME=\"/usr/lib/jvm/java-${version}-oracle\"",
}
}
Log of run:
Notice: Compiled catalog for precise64 in environment production in 0.78 seconds
Notice: /Stage[main]/Apt/File[preferences]/ensure: created
Notice: /Stage[main]/Rap::Java/Exec[accept-java-license]/returns: executed successfully
Notice: /Stage[main]/Rap::Java/File_line[java_environment]/ensure: created
Notice: /Stage[main]/Apt/Apt::Setting[conf-update-stamp]/File[/etc/apt/apt.conf.d/15update-stamp]/ensure: defined content as '{md5}0962d70c4ec78bbfa6f3544ae0c41974'
Notice: /Stage[main]/Rap::Java/Apt::Ppa[ppa:webupd8team/java]/Package[python-software-properties]/ensure: created
Notice: /Stage[main]/Rap::Java/Apt::Ppa[ppa:webupd8team/java]/Exec[add-apt-repository-ppa:webupd8team/java]/returns: executed successfully
Notice: /Stage[main]/Apt::Update/Exec[apt_update]: Triggered 'refresh' from 1 events
Notice: /Stage[main]/Java/Package[java-common]/ensure: created
Notice: /Stage[main]/Java/Package[java]/ensure: created
Notice: Applied catalog in 39.58 seconds
To extend the question further, generally things that are blockers for a standard setup to run are usually moved into a run stage (documented here).
So I would probably move all of the various repo setup puppet code into pre run stage with other prerequisites (normally you put in repo setup), the run stage will always be run first before the main stage, so you don't have to worry about explictly setting requirements that repos are setup on each package. This makes making changes to repos and prerequisites a lot easier

Resources