Fresh Puppet PE 3.8 install error

[Kamil Keski]

Clean slate install on Ubuntu 14.04 produces the following error and will not proceed.

** HTTP_PROXY= http_proxy= HTTPS_PROXY= https_proxy= curl --tlsv1 -s https://puppetmaster.local:8140
** HTTP_PROXY= http_proxy= HTTPS_PROXY= https_proxy= curl --tlsv1 -s https://puppetmaster.local:8140
** HTTP_PROXY= http_proxy= HTTPS_PROXY= https_proxy= curl --tlsv1 -s https://puppetmaster.local:8140
** HTTP_PROXY= http_proxy= HTTPS_PROXY= https_proxy= curl --tlsv1 -s https://puppetmaster.local:8140

========================================================================
** cp /opt/puppet/share/installer/install_log.lastrun.ITTPuppet.log /var/log/pe-installer
** sed "s/^\(q_.*password\)=.*/#\1=REDACTED/g" < "/tmp/pe-installer-M4MbxK1G/install/answers.lastrun.ITTPuppet" > "/var/log/pe-installer/answers.install"
!! ERROR: The puppet master service failed to start within 120
   seconds; unable to proceed

I am using puppetmaster.local as the FQDN which is resolvable from the server and the client machine I'm executing the web install from (mapped in hosts file for testing).  Nothing else of interest in the log files.  Have tried on two separate clean images to no avail.

pe-puppetserver log shows the following.

Failed to load feature test for posix: can't find user for 0

Cannot run on Microsoft Windows without the win32-process, win32-dir and win32-service gems: Win32API only supported on win32

Puppet::Error: Cannot determine basic system flavour

             (root) at /opt/puppet/lib/ruby/site_ruby/1.9.1/puppet/feature/base.rb:32

            require at org/jruby/RubyKernel.java:1071

Well that's interesting..

All default packages have been updated.

Maddening!

Any idea's?

Thanks,

K

[Josh Cooper]

--
You received this message because you are subscribed to the Google Groups "Puppet Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to puppet-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/puppet-users/75aa1033-96b8-4a09-9028-447d6e27eb1b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

I agree the error message is anything but useful, and you're not the first to run into it. I think this is the issue you're seeing https://docs.puppetlabs.com/puppetserver/latest/known_issues.html#tmp-directory-mounted-noexec

Josh

--

Josh Cooper

Developer, Puppet Labs

PuppetConf 2015 is coming to Portland, Oregon! Join us October 5-9.

Register now to take advantage of the Early Adopter discount —save $349!

[Jay Benner]

I've been chasing this around all day today.  I have no idea what this is or how to get around it.  

[Kamil Keski]

Thanks for the reply Josh.  I worked through this and other posts I found pointing to the tmp issue.  However I don't have /tmp or /usr/tmp mounted, they are just dir's on the root mount.

fstab:

# <file system> <mount point>   <type>  <options>       <dump>  <pass>

/dev/mapper/ITTPuppet--vg-root /               ext4    errors=remount-ro 0       1

# /boot was on /dev/sda1 during installation

UUID=89c2c14f-03a6-4f43-9935-25061c52e4f9 /boot           ext2    defaults        0       2

/dev/mapper/ITTPuppet--vg-swap_1 none            swap    sw              0       0

/dev/fd0        /media/floppy0  auto    rw,user,noauto,exec,utf8 0       0

Pretty vanilla stuff here, so noexecute should not be an issue in this instance.  For good measure I tried creating a new dir (/usr/tmp) with 777 and defined -Djava.io.tmpdir=/usr/tmp globally via my /bashrc profile for the user I'm running under and root.  Verified that was picked up via java -version.

Picked up _JAVA_OPTIONS: -Djava.io.tmpdir=/usr/tmp

java version "1.6.0_35"

OpenJDK Runtime Environment (IcedTea6 1.13.7) (6b35-1.13.7-1ubuntu0.12.04.2)

OpenJDK 64-Bit Server VM (build 23.25-b01, mixed mode)

Still same behavior.  I don't see how anyone has worked through a fresh install on Ubuntu at this point.

Thanks,

K

[Kamil Keski]

Picked up _JAVA_OPTIONS: -Djava.io.tmpdir=/usr/tmp

java version "1.7.0_79"

OpenJDK Runtime Environment (IcedTea 2.5.5) (7u79-2.5.5-0ubuntu0.14.04.2)

OpenJDK 64-Bit Server VM (build 24.79-b02, mixed mode)

The last printout was from another fresh image I tried on 12.04 (same result).

[Martin Alfke]

Hi Kamil,

have you tried setting the path attribute at the exec resource?
Puppet exec starts with an empty environment and you are using several Unix commands without full path.

hth,
Martin

To view this discussion on the web visit https://groups.google.com/d/msgid/puppet-users/ac8ddeaf-d59d-45d0-a83f-abc031dc12fe%40googlegroups.com.

[Kamil Keski]

Yes, I've added the temp path arg to the pe-puppetserver init as well.  No go.

Has anyone installed successfully from scratch using an ubuntu 12.04 or 14.04 distro?  Not an existing one, but a nice fresh build?

[charles crowder]

I'm having the same problem on RHEL 6.6

I opened a support case with puppet, and they had me check the usual suspects like the /tmp noexec.  I confirmed that pe-puppetserver is up and listening on 8140, and I also confirmed that I could successfully curl to the FQDN after stripping out the http_proxy variables.  

My suspicion is that the installer is either not appropriately handling those proxy variables when they are unpopulated.  Technically we don't need those variables at all, at least in my environment.  Could be that a hack to the install script will do the trick, but I'm hesitant to do that without a sign off from the support folks.

Thanks,

-Charles

[Kamil Keski]

Going to answer my own question here as there is likely no way a fresh install of PE 3.8 will succeed on Ubuntu 12.04 or 14.04.  Ruby comes preinstalled with the distro's with a version of 1.9 (old for sure).  However this should be known.  I Installed 2.1.5, made global and the installation of PE 3.8 succeeded.

I'd like to see the PE install updated to pull a version of Ruby that it requires (who knows, 2.0 may be sufficient).  At minimum listing that you have a minimum version of Ruby installed as a prerequisite.

[Kamil Keski]

Scratch 2.1.5 as the issue.  Quick lsof shows that everything is still using the 1.9.1 binary pulled in the install.  Only other thing I installed was the following packages as they are my default pulls for a ruby build.

zlib1g-dev build-essential libssl-dev libreadline-dev libyaml-dev libsqlite3-dev sqlite3 libxml2-dev libxslt1-dev libcurl4-openssl-dev python-software-properties

I'm not going any deeper than that on this one.  I'm tired of drudging through .rb's and am leaving this as is.

Still, on a pure fresh Ubuntu 12.04 or 14.04 install there is a dependency that the PE installer is not pulling down.

If you're a user and need a quick fix just install the above packages.

Puppet devs, if you can track down which and package that into the installer that would be fantastic.

Thanks,

K

[Jay Benner]

Nailed it!  Not sure which package did the trick, but smooth sailing after installing these.

[Martin Alfke]

Confirmed.
This package list also solves installation issue on Ubuntu 14.10

Many thanks,

Martin

> To view this discussion on the web visit https://groups.google.com/d/msgid/puppet-users/618cd0ca-0687-4990-a5c8-646f9ef402f7%40googlegroups.com.

[Martin Alfke]

I have created a blog posting with PE 3.8 installation issues and fixes on Debian 6 and 7, CentOS 6 and 7, SLES 11 and 12, Ubuntu 12.04 and 14.10:
http://tuxmea.blogspot.de/2015/05/puppet-enterprise-38-installation-steps.html

[Jeremy Barlow]

We found a likely explanation for the Puppet Server service not starting on PE 3.8.

The newer Puppet Server version in PE 3.8, through transitive dependencies, needs to be able to load the "libcrypt.so" shared library.  On some Ubuntu configurations, the LD_LIBRARY_PATH doesn't include the directory that this library lives in.  We found this to be the case with with a default AMI amd64 Ubuntu 14.04 in Amazon EC2, for example.  This results in the master failing to start with this cryptic error message.

One can workaround the problem by updating the library path that Puppet Server uses at startup.  Instructions on how to do that for PE are now documented here - http://docs.puppetlabs.com/pe/latest/release_notes_known_issues.html#ubuntu-conflict-when-youre-running-on-fusion-vm-and-amazon-ec2.  Also, we have a JIRA ticket with more information about this problem that we'll use for tracking the longer-term fix - see https://tickets.puppetlabs.com/browse/SERVER-592.

We would be interested in learning if there are any users who continue to encounter this problem after trying the workaround.

--- Jeremy

[Paolo Marin]

I am having the same issue. 

Installing PE 3.8 to ubuntu 12.04.

Installer fail with the following:

" ** cp /opt/puppet/share/installer/install_log.lastrun."FQDN".log /var/log/pe-installer

** sed "s/^\(q_.*password\)=.*/#\1=REDACTED/g" < "/tmp/pe-installer-wU0wqDPV/install/answers.lastrun."FQDN" > "/var/log/pe-installer/answers.install"
!! ERROR: The puppet master service failed to start within 120
   seconds; unable to proceed

I have tried 3 times and I thing this is a bug with the product.

Please advice of any solution.

Thanks

[Martin Alfke]

Hi Paolo,

the answer was a few mails earlier:

install the following package list:

zlib1g-dev build-essential libssl-dev libreadline-dev libyaml-dev libsqlite3-dev sqlite3 libxml2-dev libxslt1-dev libcurl4-openssl-dev python-software-properties

Also see http://tuxmea.blogspot.de/2015/05/puppet-enterprise-38-installation-steps.html for other Ubuntu related installation issues.

Best,

Martin

> To view this discussion on the web visit https://groups.google.com/d/msgid/puppet-users/6d3661a1-66a4-4e86-9242-dad25fb2f1ec%40googlegroups.com.

[Andrew Coggins]

The above steps fixed my installer issues on Ubuntu 14.04 running on a VirtualBox VM, however moving to AWS and going through the steps again, I can't successfully install puppet 3.8.1.

The unusual thing is I don't see any issues logged in puppetserver.log and telnet to localhost 8140 works. Any ideas?

Thanks

[Andrew Coggins]

Fixed my issue. I fixed up the hosts file according to this documentation: http://docs.puppetlabs.com/pe/latest/release_notes_known_issues.html#debianubuntu-local-hostname-issue

I had puppet pointing to 127.0.0.1 in the hosts file. Changing this to the private IP of the ec2 instance resolved the issue.