Home / user-29328

raphink's questions

Martin Hope
raphink
Asked: 2014-03-18 08:59:24 +0800 CST
  • 0

Using MCollective (2.2.3) with RabbitMQ (3.1.5) on RedHat 6.5, MCollective stops responding after a first successful ping.

When I restart RabbitMQ, it works again.

root@rabbitmqserver# /etc/init.d/rabbitmq-server restart
Restarting rabbitmq-server: SUCCESS
rabbitmq-server.
root@rabbitmqserver# rabbitmqctl list_consumers -p /mcollective  |  grep directed
amq.gen-1amX9Z9w6jBKNP7IQrMfjg  <[email protected]>  Q_/exchange/mcollective_directed/srv114 false
amq.gen-dxFvTuxvQZOU0d1e7XuJJg  <[email protected]>  Q_/exchange/mcollective_directed/srv98  false
amq.gen-lwttnRCCTON9hgPg7X2zZw  <[email protected]> Q_/exchange/mcollective_directed/srv99  false
amq.gen-pG4o5csOGcam9mxvIQvQ7g  <[email protected]>  Q_/exchange/mcollective_directed/srv57  false
amq.gen-xtRFm6aAoeV2mDO9dTAyyA  <[email protected]>  Q_/exchange/mcollective_directed/srv115 false


user@mco-client$ mco ping

srv114                        time=66.49 ms
srv57                         time=69.34 ms

---- ping statistics ----
2 replies max: 69.34 min: 66.49 avg: 67.91 
user@mco-client$  mco ping


---- ping statistics ----
No responses received

root@rabbitmqserver# rabbitmqctl list_consumers -p /mcollective  |  grep directed
amq.gen-1amX9Z9w6jBKNP7IQrMfjg  <[email protected]>  Q_/exchange/mcollective_directed/srv114 false
amq.gen-dxFvTuxvQZOU0d1e7XuJJg  <[email protected]>  Q_/exchange/mcollective_directed/srv98  false
amq.gen-lwttnRCCTON9hgPg7X2zZw  <[email protected]> Q_/exchange/mcollective_directed/srv99  false
amq.gen-pG4o5csOGcam9mxvIQvQ7g  <[email protected]>  Q_/exchange/mcollective_directed/srv57  false
amq.gen-xtRFm6aAoeV2mDO9dTAyyA  <[email protected]>  Q_/exchange/mcollective_directed/srv115 false

If I try using another command (such as package), it fails right after the discovery (after restarting RabbitMQ that is):

user@mco-client$ mco package status bash
Do you really want to operate on packages unfiltered? (y/n): y

 | [ >                                                             ] 0 / 2
[... stalled ...]

After a while, I start seeing this in my MCollective server logs:

E, [2014-03-18T11:51:01.332969 #24456] ERROR -- : base.rb:29:in `run' Sending registration message failed: Stomp::Error::NoCurrentConnection
E, [2014-03-18T11:55:35.759653 #24456] ERROR -- : agents.rb:138:in `dispatch' Execution of discovery failed: Stomp::Error::NoCurrentConnection
E, [2014-03-18T11:55:35.760040 #24456] ERROR -- : agents.rb:139:in `dispatch' /usr/lib/ruby/gems/1.8/gems/stomp-1.2.2/lib/stomp/connection.rb:331:in `publish'
        /usr/libexec/mcollective/mcollective/connector/rabbitmq.rb:184:in `publish'
        /usr/lib/ruby/site_ruby/1.8/mcollective/message.rb:230:in `publish'
        /usr/lib/ruby/1.8/timeout.rb:67:in `timeout'
        /usr/lib/ruby/site_ruby/1.8/mcollective/message.rb:218:in `publish'
        /usr/lib/ruby/site_ruby/1.8/mcollective/runner.rb:129:in `reply'
        /usr/lib/ruby/site_ruby/1.8/mcollective/runner.rb:83:in `agentmsg'
        /usr/lib/ruby/site_ruby/1.8/mcollective/agents.rb:132:in `dispatch'
        /usr/lib/ruby/1.8/timeout.rb:67:in `timeout'
        /usr/lib/ruby/site_ruby/1.8/mcollective/agents.rb:125:in `dispatch'
        /usr/lib/ruby/site_ruby/1.8/mcollective/agents.rb:121:in `initialize'
        /usr/lib/ruby/site_ruby/1.8/mcollective/agents.rb:121:in `new'
        /usr/lib/ruby/site_ruby/1.8/mcollective/agents.rb:121:in `dispatch'
        /usr/lib/ruby/site_ruby/1.8/mcollective/runner.rb:82:in `agentmsg'
        /usr/lib/ruby/site_ruby/1.8/mcollective/runner.rb:55:in `run'
        /usr/lib/ruby/site_ruby/1.8/mcollective/runner.rb:50:in `loop'
        /usr/lib/ruby/site_ruby/1.8/mcollective/runner.rb:50:in `run'
        /usr/lib/ruby/site_ruby/1.8/mcollective/unix_daemon.rb:30:in `daemonize_runner'
        /usr/lib/ruby/site_ruby/1.8/mcollective/unix_daemon.rb:13:in `daemonize'
        /usr/lib/ruby/site_ruby/1.8/mcollective/unix_daemon.rb:5:in `fork'
        /usr/lib/ruby/site_ruby/1.8/mcollective/unix_daemon.rb:5:in `daemonize'
        /usr/lib/ruby/site_ruby/1.8/mcollective/unix_daemon.rb:20:in `daemonize_runner'
        /usr/sbin/mcollectived:43
E, [2014-03-18T11:56:01.336954 #24456] ERROR -- : base.rb:29:in `run' Sending registration message failed: Stomp::Error::NoCurrentConnection
E, [2014-03-18T12:01:01.343415 #24456] ERROR -- : base.rb:29:in `run' Sending registration message failed: Stomp::Error::NoCurrentConnection
E, [2014-03-18T12:06:01.347218 #24456] ERROR -- : base.rb:29:in `run' Sending registration message failed: Stomp::Error::NoCurrentConnection
E, [2014-03-18T12:11:01.354069 #24456] ERROR -- : base.rb:29:in `run' Sending registration message failed: Stomp::Error::NoCurrentConnection

even though the machine is still listed as a consumer on the RabbitMQ server.

What could be going wrong here?

rabbitmq
Martin Hope
raphink
Asked: 2013-12-19 03:11:48 +0800 CST
  • 1

I'm trying to set up a fully redundant OpenVPN tunnel, with two routers on each side.

All the tutorials I saw (such as http://dev.vyatta4people.org/highly-available-openvpn-connection-between-two-offices/) mention a site-to-site setting, which doesn't seem to work with OpenVPN on a Linux server (which seems to always require a server and a client).

How can I achieve a similar setup using Linux servers?

networking
Martin Hope
raphink
Asked: 2012-11-13 06:25:17 +0800 CST
  • 1

I'm setting up Puppetdb with SSL and having issues with certificates.

I'm using Nginx as an SSL proxy for Puppet, so my CA is managed by a mongrel server on this Nginx proxy machine.

If I generate a certificate for my Puppetdb URI using the CA on the Nginx machine, I'm able to setup the Puppetdb using the puppetlabs-puppetdb module (since the Puppet agent uses the proxy's CA), but then the Puppetmaster cannot connect to it, because it has its own CA certificate which is generates itself.

If I generate a certificate for the Puppetdb URI using one of the Puppetmasters, I cannot deploy the Puppetdb using the puppetlabs-puppetdb module since the Puppet agent doesn't use the same CA certificate.

What could I do to reconcile all this? Can I full turn off SSL on my puppetmasters (since SSL is managed by the Nginx proxy) and have them use the proxy's CA to connect to the Puppetdb?

nginx
Martin Hope
raphink
Asked: 2012-10-20 02:25:43 +0800 CST
  • 4

I have a firewall that has 3 IP aliases on 1 physical interface. Packets get dropped between these 3 interfaces (either ICMP, HTTP, or anything else). We tracked it down to these packets being marked INVALID in the FORWARD rule and dropped due to the this rule:

chain FORWARD {
    policy DROP;

    # connection tracking
    mod state state INVALID LOG log-prefix 'INVALID FORWARD DROP: '; 
    mod state state INVALID DROP;
    mod state state (ESTABLISHED RELATED) ACCEPT;
}

(That is, we see the INVALID FORWARD DROP logs in dmesg)

What could be causing this?

iptables
Martin Hope
raphink
Asked: 2012-04-20 21:44:58 +0800 CST
  • 2

In puppet, modules can have a dash in their name, and thus variables can, too. But then, how can we use such variables?

For example, I have a $ssh_mode variable defined in a google-authenticator::params class, so I expect to be able to call $google-authenticator::params::ssh_mode, but puppet (0.25.5) cuts on the dash. Using ${google-authenticator::params::ssh_mode} doesn't help either.

What can I do to access this variable?

puppet
Martin Hope
raphink
Asked: 2012-04-05 05:30:48 +0800 CST
  • 0

I'm trying to add a require to the default resource for users in Puppet:

User { require => Mount['/home'] }

Using Puppet 0.25, I get the following error:

err: Could not retrieve catalog from remote server: Error 400 on SERVER: Syntax error at '+>'; expected '}' at /home/rpinson/puppetmaster/manifests/foo.pp:71 on node ip-bar.internal

Is there a way to achieve that?

puppet
Martin Hope
raphink
Asked: 2012-03-21 12:42:48 +0800 CST
  • 4

I'm trying to think of a clean way to determine the location of machines (mainly, which datacenter they belong to) based on their network settings.

I would like it to be dynamic, and I'm thinking of using special DNS records that would be specific to the DNS server in each datacenter.

For example, you could have:

root@machine1# dig TXT mysite
...
mysite      3600    IN  TXT "DC1"
...
root@machine2# dig TXT mysite
...
mysite      3600    IN  TXT "DC2"
...

etc.

I know that DNS has a special LOC record for location, but it takes coordinates, so it doesn't help in my case. Is there a standard way of addressing this issue, another special type of record for it, or some standard entries in TXT records?

domain-name-system
Martin Hope
raphink
Asked: 2010-10-07 07:50:46 +0800 CST
  • 0

I often find myself having to clean up some directories because they are too full. Let's take /var as an example, as it is a common one, for obvious reasons.

In order to see which sub-directories takes a lot of space, I use du -sh /var/*. Unfortunately, I have some huge data in sub-directories of /var, which happen to be hosted on other partitions than /var (e.g. /var/chroot) and greatly slow the process for nothing.

Now du has a -x flag which excludes other mountpoints. It works fine when doing a du -shx /var but not with du -shx /var/* since * is expanded by the shell so /var/chroot is explicitely sent to du.

Another option I was proposed was to use find /var -xdev -mindepth 1 -maxdepth 1 -exec du -sh {} \+ but again, even though find does not follow other mountpoints, it still prints them and passes them to du, hence failing again to achieve my goal.

Before I end up writing a horrible bash function (or a perl/python program) that parses df or mount in a loop, does anyone know of a clean way to achieve what I'm trying to do?

linux unix shell directory