Ask Your Question
3

unable to ping metadata server 169.254.169.254

asked 2013-05-10 08:16:54 -0500

anonymous user

Anonymous

updated 2013-08-22 19:41:18 -0500

fifieldt gravatar image

We are able to create a VM with demo-net IP 10.5.5.2 and attached a floating IP 7.7.7.131 and was able to ping but still facing few issues.

  • Unable to connect it through SSH. We have executed the following command

    ssh -i keypair.pem root@10.5.5.2 Read from socket failed: Connection reset by peer

  • In the instance console log we can see lots of warning like the following.

    20130421 14:43:52,548 util.py[WARNING]: 'http://169.254.169.254/20090404/metadata/instanceid' failed [0/120s]: http error [500]

IP 169.254.169.254 is not pingable , using route -n we could see an entry for 169.254.0.0. We tried to remove this entry using route del but again it reappear in next network restart. We are using ubuntu 12.04 LTS.

  • From horizon dashboard we can see a console asking for login in the following path instance - console. But we do not have the proper loginid and password for the created instance. We tried several combinations like (admin, blank) (root, blank) (ubuntu, blank) (demo, passoword) (admin, password) but none of them are working.
edit retag flag offensive close merge delete
add a comment

5 answers

Sort by ยป oldest newest most voted
3

answered 2013-05-10 10:07:45 -0500

darragh-oreilly gravatar image

updated 2013-05-10 10:26:10 -0500

  1. If the metadata service cannot be reached from instance, then it won't have the public key - try ssh without -i keypair.pem
  2. I doubt the meta service is suppose to respond to icmp.
  3. If you are using the cirros test image, then the username is cirros and the password is cubswin:)
edit flag offensive delete link more

Comments

Thanks for your input Darragh. Currently we are able to login to the console from horizon dashboard using cirros image.

But still the instances are not able to access the metadata server 169.254.169.254. While booting we are getting lots of error messages like

========================================================

cloudsetup: checking http://169.254.169.254/20090404/metadata/instanceid wget: server return error: HTTP/1.1 500 Internal Server Error cloudsetup: failed 1/30: up 13.01. request failed wget: server return error: HTTP/1.1 500 Internal Server Error cloudsetup: failed 2/30: up 14.42. request failed

cloudsetup: after 30 fails , debugging cloudsetup: running debug (30 tries reached)

========================================================

Please help us to resolve this.

sourav gravatar imagesourav ( 2013-05-13 07:28:19 -0500 )edit
1

What networking are you using (nova-network or quantum)? What is the OpenStack version?

darragh-oreilly gravatar imagedarragh-oreilly ( 2013-05-13 12:00:35 -0500 )edit

We are using quantum. We are not sure about openstack version kindly let us know how to check that. "nova-manage --version list" shows the output as 2013.1

sourav gravatar imagesourav ( 2013-05-15 03:38:42 -0500 )edit

That would be Grizzly. In Grizzly there are two ways for the VM to reach the metadata server. One via a Quantum router and the other with DHCP. I haven't tried the DHCP way, but I believe it injects a route for 169.254 into the VM, but I don't know how it works after that - sorry.

darragh-oreilly gravatar imagedarragh-oreilly ( 2013-05-15 04:25:10 -0500 )edit
1

Salvatore Orlando explains how it works here http://www.mail-archive.com/openstack@lists.launchpad.net/msg22432.html

darragh-oreilly gravatar imagedarragh-oreilly ( 2013-05-15 05:23:51 -0500 )edit
add a comment
0

answered 2013-07-27 13:27:20 -0500

racedo gravatar image

Check if you are hitting this bug https://bugs.launchpad.net/nova/+bug/1184641

edit flag offensive delete link more
add a comment
0

answered 2013-05-27 02:10:05 -0500

openstackpva gravatar image

I am also unable to ping instance "i-00000001" with ip 172.16.0.1 from client machine although I can ping from client(172.16.0.22)to server(172.16.0.21). All the services are running on server nova-compute, nova-network, nova-cert, nova-scheduler, sshd, ports enabled. I am getting error message: ssh: connect to host 172.16.0.1 port 22: No route to host

What to do now?

edit flag offensive delete link more
add a comment
0

answered 2014-09-24 04:25:10 -0500

zmask gravatar image

Guys even i am facing the same issue. Has this issue been resolved. ?

edit flag offensive delete link more
add a comment
0

answered 2013-05-13 08:06:59 -0500

Alen Komljen gravatar image

To ssh to Ubuntu machine use ubuntu user:

ssh ubuntu@10.5.5.2 -i keypair.pem

Metadata will not respond to ping, if you want to check that try:

curl http://169.254.169.254/
edit flag offensive delete link more

Comments

We were unable to connect the instance using ubuntu username as well, but currently we are able to login using cirros test image. The metadata server 169.254.169.254 is still not accessible, curl http://169.254.169.254 shows the output as "Curl: (7) couldn't connect to host"

sourav gravatar imagesourav ( 2013-05-15 03:44:33 -0500 )edit
1

On your compute node in nova.conf do you have this property: enabled_apis=metadata?

Alen Komljen gravatar imageAlen Komljen ( 2013-05-15 05:40:35 -0500 )edit
add a comment

Get to know Ask OpenStack

Resources for moderators

Question Tools

2 followers

subscribe to rss feed

Stats

Asked: 2013-05-10 08:16:54 -0500

Seen: 10,437 times

Last updated: Sep 24 '14

Related questions

Can't ping virtual machines! [closed]

Basic active/standby HA scenario issue

Cant ping self-service network [closed]

Self-service User Password Change - NotAuthorized

How to change hostname/local-hostname/public-hostname in metadata

liberty install on ubuntu trouble with self-service networs

Update config-drive after metadata change?

Admin password from create server api response is not working

Cant ping floating ip

LXC instance booted and active but can not be pinged