[Openstack] glance access error from remote client
Hello, I have an Essex install of openstack on RHEL6. The controller node has all openstack services running, and there are several separate compute nodes. Now I have a problem using glance. If I issue glance index --host=XXX --port=9292 on the controller node, it works fine. If I issue the same command from a remote client node, where I have all the relevant openstack rpms installed, it fails with the following error message: Failed to show index. Got error: There was an error connecting to a server Details: [Errno 111] Connection refused On the same client node, I can query images and instances using euca2ools, they just work fine. And I also can telnet to the 9292 port on the controller node from the client side. In both cases, I use the credentials of the admin account. Any wisdom on what could cause this problem? Thanks, Xin smime.p7s Description: S/MIME Cryptographic Signature ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] glance access error from remote client
On 06/27/2012 10:04 AM, Xin Zhao wrote: Hello, I have an Essex install of openstack on RHEL6. The controller node has all openstack services running, and there are several separate compute nodes. Now I have a problem using glance. If I issue glance index --host=XXX --port=9292 on the controller node, it works fine. If I issue the same command from a remote client node, where I have all the relevant openstack rpms installed, it fails with the following error message: Failed to show index. Got error: There was an error connecting to a server Details: [Errno 111] Connection refused This usually because the glance CLI tool is trying to connect to the localhost (which obviously isn't running the Glance API server). Have you looked at the difference between your OS_XXX environment variables between the controller node and the remote client node? It may be that when you are doing a glance index --host=XXX --port=9292 on the controller node that the glance CLI tool is actually connecting to localhost/127.0.0.1, but I'm not completely sure.. Best, -jay ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] glance access error from remote client
Hi Jay, I thought about that, and changed the client environment variables to use the FQDN of the controller, I even restarted glance/swift services with the controller FQDN in the config files (127.0.0.1/localhost are the original values). I compared the OS_XXX env between the controller node and remote client, they are the same. What's interesting is that, even if I give a wrong hostname or port number in the command glance index --host=XXX --port=XXX, it still works if run on the controller node. And, strace shows, from the remote client node, that 127.0.0.1:9292 is connected instead of the hostname:9292 specified in the command line. So how can I make sure the client does talk to THE hostname, instead of 127.0.0.1 ? Thanks, Xin On 6/27/2012 11:40 AM, Jay Pipes wrote: On 06/27/2012 10:04 AM, Xin Zhao wrote: Hello, I have an Essex install of openstack on RHEL6. The controller node has all openstack services running, and there are several separate compute nodes. Now I have a problem using glance. If I issue glance index --host=XXX --port=9292 on the controller node, it works fine. If I issue the same command from a remote client node, where I have all the relevant openstack rpms installed, it fails with the following error message: Failed to show index. Got error: There was an error connecting to a server Details: [Errno 111] Connection refused This usually because the glance CLI tool is trying to connect to the localhost (which obviously isn't running the Glance API server). Have you looked at the difference between your OS_XXX environment variables between the controller node and the remote client node? It may be that when you are doing a glance index --host=XXX --port=9292 on the controller node that the glance CLI tool is actually connecting to localhost/127.0.0.1, but I'm not completely sure.. Best, -jay ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp smime.p7s Description: S/MIME Cryptographic Signature ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp