Ask Your Question

nosluoc's profile - activity

2014-01-23 02:21:17 -0500 received badge  Good Answer (source)
2013-12-15 10:03:17 -0500 received badge  Nice Answer (source)
2013-12-13 10:12:06 -0500 received badge  Commentator
2013-12-13 10:12:06 -0500 commented question nova-compute not update vm status when compute node restarted or shutdown

Bumping for a legitimate question-- same issue encountered in my cluster.

2013-08-29 17:10:52 -0500 received badge  Student (source)
2013-08-21 14:07:49 -0500 answered a question ERROR [cinder.openstack.common.rpc.common] AMQP server on localhost:5672 is unreachable: Socket closed.

I ran into this problem when installing/configuring DevStack where just like you, the port was open, I verified the RabbitMQ service was running, AND checked my credentials. I stumbled upon this page, which had me change the RabbitMQ guest user password, and restart the RabbitMQ service, which remedied my problem. Not entirely sure why this would have happened, but it might help you.

Change the RabbitMQ guest password: sudo rabbitmqctl change_password guest password

Restart RabbitMQ service: sudo service rabbitmq restart

2013-08-19 17:00:01 -0500 commented answer Looking for an OpenStack monitoring tool

Link is broken-- some added characters at the end.

2013-08-19 16:58:43 -0500 answered a question Looking for an OpenStack monitoring tool

Ceilometer is a somewhat obvious choice simply because it's natively integrated with OpenStack. I know that Puppet Labs were planning to find a way to implement Nagios/Icinga/Sensu/etc monitoring, but don't think that's a feature they've added just yet.

Reference: http://projects.puppetlabs.com/issues/14341

2013-08-16 04:28:10 -0500 received badge  Famous Question (source)
2013-08-12 07:10:28 -0500 received badge  Notable Question (source)
2013-08-01 13:12:04 -0500 commented question When I use swift to perform a HEAD, I am getting 401 Unauthorized.

Jimmy, I don't have a specific answer for this question right now, however, this issue appears to be related, and the OP mentioned he resolved the issue, but hasn't yet responded with how he did so: https://ask.openstack.org/question/1941/swift-error-unable-to-retrieve-container-list/

2013-08-01 12:09:07 -0500 commented answer Is there a "max" number of zones in Swift?

None. It's entirely up to the admin to define the number/scope of zones.

2013-08-01 12:05:44 -0500 commented answer nova-client: ERROR: Unauthorized (HTTP 401)

I'm not 100% positive on this, so hopefully someone can come by to confirm/correct me if I'm wrong, but: I'm fairly sure that in this case, the *.ini files are utilized because when Nova creates a VM, it uses libvirt to authenticate, which requires a *.ini file/syntax. http://libvirt.org/auth.html

2013-08-01 11:52:38 -0500 answered a question In cielometer : For network statistics like network.incoming.bytes, and network.outgoing.bytes, is it an aggregate of bytes on all the interfaces of a VM, please clarify

Yes. Both 'network.incoming.bytes' and 'network.outgoing.bytes' are cumulative/aggregate values representing incoming/outgoing traffic on a VM's interface.

Source: OpenStack Documentation - Ceilometer Measurements

2013-08-01 09:40:09 -0500 commented answer Is there a "max" number of zones in Swift?

Thank you for converting the URL to a hyperlink. :)

2013-08-01 09:38:08 -0500 received badge  Supporter (source)
2013-08-01 09:33:44 -0500 commented answer High-level Storage Question

Yes, I'm aware of the danger of using two nodes (split brain in the event of failure). We're planning to re-tool a few spare storage servers we have sitting around, so for all intents and purposes, I probably could have just left out the part stating that we're only using two servers..

2013-07-31 16:59:00 -0500 received badge  Teacher (source)
2013-07-31 15:10:37 -0500 received badge  Editor (source)
2013-07-31 15:09:57 -0500 answered a question Is there a "max" number of zones in Swift?

According to OpenStack's Swift Deployment Guide (http://docs.openstack.org/developer/swift/deployment_guide.html), zones are 'abstract' in that you define how much separation exists between the zones in your deployment. To my knowledge, there is no limit to the number of zones you can have. It's more important to plan the number of zones according to your application: are all zones powered separately? Are they on the same switch? As far as Swift is concerned, zones are used to isolate data and to clearly define failure boundaries. You can define a zone as a single drive at the smallest level, up to multiple racks on a larger scale. This also gives Swift a sense of the health/status of the replicas, allowing it to tolerate outages of zones without losing all available replicas of data.

TL;DR - Since zones are relative to the use case, there's no definite answer to your question. A zone could be a single drive that fails, or it could be an entire cabinet. Generally speaking, if properly planned out, more zones would increase the fault tolerance of Swift.

2013-07-31 13:25:33 -0500 commented answer Why the changes to keystone.conf gets back to same when I run ./stack.sh in devstack ?

I remember reading that it does this by design to prevent individuals from using DevStack in production environments. Can't quote a source at the moment, but makes sense to me.

2013-07-31 13:01:07 -0500 commented question Where Can I download slide from Openstack Havana Submit

It would probably be easier if you specified which workshop/presentation you're looking for first, and then maybe someone can link you to either Slideshare or YouTube..

2013-07-31 12:53:35 -0500 received badge  Popular Question (source)
2013-07-31 09:12:48 -0500 commented answer High-level Storage Question

Thanks for your reply! I did see your response to that question, however, it seemed highly tailored towards Cinder and Glance, with no real mention on integrating Ceph with "Swift".

2013-07-30 15:33:16 -0500 asked a question High-level Storage Question

Just looking for some confirmation and/or correction: from reading the OpenStack manuals and all available information on Ceph, it's my understanding that Ceph can act as a backend for Swift/Cinder/Glance, is this correct?

I'm asking because we're in the process of planning our OS deployment, and we have two relatively large storage servers (we plan to add more in the very near future), and would like to use Ceph to avoid dedicating one server to Swift, and the second to Cinder with no redundancy. Any additional thoughts/input/better ideas are more than welcomed.. thanks in advance!