ec2_facts hung

[Steven Truong]

Hi all,

The subject is misleading but I would like to get some helps on how to troubleshoot this or figure out what can be done.

I have about 70 instances with tag kafkaproducer=yes and I want to do rolling updates to them but my playbook went nowhere with ec2_facts.  

I ran the module ec2_facts on the command line and at around 37 instances, the thing just stood there for eternity.

ansible tag_kafkaproducer_yes -i ec2.py   -m ec2_facts -vvvv -f 10

.....

<ec2-.......compute-1.amazonaws.com> EXEC ['ssh', '-C', '-tt', '-vvv', '-o', 'ControlMaster=auto', '-o', 'ControlPersist=60s', '-o', 'ControlPath=/home/steven/.ansible/cp/ansible-ssh-%h-%p-%r', '-o', 'IdentityFile="/home/steven/.ssh/key-ec2"', '-o', 'KbdInteractiveAuthentication=no', '-o', 'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', '-o', 'PasswordAuthentication=no', '-o', 'User=ec2-user', '-o', 'ConnectTimeout=10', u'ec2......compute-1.amazonaws.com', u'/bin/sh -c \'su root -c "/bin/sh -c \'"\'"\'echo SUDO-SUCCESS-zfbckatzfxprscqyvqbrpepuokyhmtnn; LANG=en_US.UTF-8 LC_CTYPE=en_US.UTF-8 /usr/bin/python /home/ec2-user/.ansible/tmp/ansible-tmp-1410222363.53-10317830534118/ec2_facts; rm -rf /home/ec2-user/.ansible/tmp/ansible-tmp-1410222363.53-10317830534118/ >/dev/null 2>&1\'"\'"\'"\'']

----

I am pretty sure that I have a recent version of OpenSSH on my servers but if this does not work then I certainly will not be able to do much things with AWS using Ansible.

Thanks,

Steven.

[Michael DeHaan]

Hmm, that's interesting.

This may not help, but what version of Ansible are you on and is this the first task in your playbook?  (Also, is gather_facts enabled?  If it gathers facts, this is the 2nd task and my question is not relevant).

--
You received this message because you are subscribed to the Google Groups "Ansible Project" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ansible-proje...@googlegroups.com.
To post to this group, send email to ansible...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ansible-project/f0254697-6a39-41e8-9d3f-5e35f6ae9dae%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Steven Truong]

Hi Michael,

I am using 1.7.1 and gather_facts was enabled by default.

And just this morning, I tried again then things are working just fine.  At this point, it is not an issue but it did happen for me.

Sorry for wasting your times, Michael.

Thanks for your helps,

Steven.

[Michael DeHaan]

Nah, it's ok.

I'm wondering if the EC2 metadata service had a glitch and was just giving you a very long timeout?

To view this discussion on the web visit https://groups.google.com/d/msgid/ansible-project/115c241f-0724-4858-9a09-6d14d862a9fb%40googlegroups.com.