Re: [Openstack] ML2 Plugin and vif_type=binding_failed

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi Raphael, please check if your ovs_plugin config is the same as the ml2 config. In addition i'm missing in your nova.conf: libvirt_vif_driver=nova.virt.libvirt.vif.LibvirtHybridOVSBridgeDriver Cheers Heiko On 20.06.2014 20:23, Raphael Ribeiro w

Re: [Openstack] ML2 Plugin and vif_type=binding_failed

Raphael, Please check if your instance is created successfully. Sometimes the instance is failed to spawn for some other reason and nova will try to clean up the instance to roll back. During the clean-up it's possible to get the vif_binding exception. You may double check your nova compute and no

[Openstack] What is the hostId in parameter in server details?

Hi, I was trying to understand how the hostId parameter in server is generated, or if is a unique id for a tenant and an user. I found this link [1] in AskOpenstack that explain better my question, and thi

[Openstack] [openstack][openstack-operators] Mysql 5.5 to 5.6

Hi, all our OpenStack databases are running on mysql 5.5.30 and we plan upgrade to 5.6.17 soon. Should we expect any issue moving to 5.6? I believe in gate 5.6 is still not used. Please correct me if I'm wrong. Profiling we see some performance decrease using 5.6 (we believe it's related with the

[Openstack] Why doesn't suspend release vCPUs/memory?

Per the Openstack documentation, it's implied that both vCPUs and memory are to be released when an instance is successfully suspended. However, when I suspend an instance, neither resources are reclaimed. Documentation: http://docs.openstack.org/user-guide/content/suspend_resume.html Administrat

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

On Mon, Jun 23, 2014 at 7:45 AM, Ricky Saltzer wrote: > > https://ask.openstack.org/en/question/32826/why-doesnt-suspend-release-vcpusmemory/ ​My understanding was always that the instance is no longer consuming any resources via the virt layer, so in essence the resources are in fact freed up

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

On Mon, Jun 23, 2014 at 10:49 AM, Ricky Saltzer wrote: > Right, the quotas don't seem to be released. If I have 210/210 vCPUs used, > and I suspend an instance with 4 vCPUs, I still have 210/210 vCPUs used. > > > On Mon, Jun 23, 2014 at 11:38 AM, John Griffith < > john.griff...@solidfire.com> wro

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

Right, the quotas don't seem to be released. If I have 210/210 vCPUs used, and I suspend an instance with 4 vCPUs, I still have 210/210 vCPUs used. On Mon, Jun 23, 2014 at 11:38 AM, John Griffith wrote: > > On Mon, Jun 23, 2014 at 7:45 AM, Ricky Saltzer wrote: > >> >> https://ask.openstack.org

Re: [Openstack] neutron, nova and vif_plugging_* in nova.conf

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 06/22/2014 12:36 AM, Gary Kotton wrote: > Hi, The flag makes the libvirt driver use the instance event > mechanism to wait for neutron to confirm that the VIF plugging is > complete before actually starting the VM. In order for this to work > correc

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

That seems to be the case, and I can see where you're coming from, but if the resources aren't released at the quota level, then they're effectively being used from a user's point of view. It would be nice if *suspend* released resources after the instance is shutdown, and a *resume* would reclaim

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

Maybe seperate quotas for active vs suspended? ​ That sounds overly complicated. At the risk of demonstrating some ignroance on my part, w have to keep in mind it isn't just quota for the tenant, but actual available resource on the compute node yes? If suspended instances are going to be

[Openstack] IPAM

Hello, I'm new to OpenStack, and I was wondering what the community is using for IPAM solutions. It seems to me that neutron is doing something of these things, but how is IPAM being handled for underlay devices such as routers and switches? Jeff d'Ambly Principal Software Engineer, Symantec C

Re: [Openstack] neutron, nova and vif_plugging_* in nova.conf

Hi, Controller node only. Thanks Gary On 6/23/14, 7:59 PM, "Dmitry S. Makovey" wrote: >-BEGIN PGP SIGNED MESSAGE- >Hash: SHA1 > >On 06/22/2014 12:36 AM, Gary Kotton wrote: >> Hi, The flag makes the libvirt driver use the instance event >> mechanism to wait for neutron to confirm that the

Re: [Openstack] neutron, nova and vif_plugging_* in nova.conf

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 06/23/2014 11:38 AM, Gary Kotton wrote: > Hi, Controller node only. Thanks! -BEGIN PGP SIGNATURE- Version: GnuPG v1 Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iD8DBQFTqGrvyDrVuGfS98QRAruiAJ0YMobG0XsPr0xqNTYr0DPCBLgoN

[Openstack] Cells options and configuration

Hello all, I wonder if anyone has implemented Cell option in your production / PoC, I checked and looks like it’s still considered experimental. Any tips on the configuration and things to be aware that will be greatly appreciated. Thanks Remo __

Re: [Openstack] Cells options and configuration

We are using cells at CERN. Currently enabling our 6th cell. There are some restrictions to bear in mind. See http://openstack-in-production.blogspot.fr/ and Belmiro's talk at the Hong Kong summit http://www.openstack.org/summit/openstack-summit-hong-kong-2013/session-videos/presentation/deep-

Re: [Openstack] IPAM

Hi Jeff, in a nutshell, Neutron has its IPAM logic baked into the main 'db class'. However, Neutron's IPAM does not manage at all the underlay - it manages exclusively devices in the 'logical' realm. There are been discussions in the past concerning Physical appliance management in Neutron, not j

Re: [Openstack] IPAM

Neutron uses an overlay model with respect to differentiating against physical switch logic. Big Switch is getting into a more literal interpretation of SDN with the control/slave model but that's an approach that's been so disruptive to implement it hasn't gained significant traction last I heard.

Re: [Openstack] Renewing keystone signing certs

On 06/23/2014 02:38 AM, Sam Morrison wrote: OK this is looking really scary, (unless I’m missing something…..) In havana and icehouse, services will only download the signing cert if it doesn’t exist. So if you replace it on the keystone server it will continue to use the existing cert. Ideal

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

Agreed, I think separating active vs suspended resources would unnecessarily complicate things. The idea here is that if you are suspending an instance, the resources that instance was using should be released back to the tenant for use. If there are no more resources available, then resuming an in

Re: [Openstack] Renewing keystone signing certs

On 06/23/2014 03:28 PM, Adam Young wrote: On 06/23/2014 02:38 AM, Sam Morrison wrote: OK this is looking really scary, (unless I’m missing something…..) In havana and icehouse, services will only download the signing cert if it doesn’t exist. So if you replace it on the keystone server it wil

[Openstack] PTL Webinars: Icehouse to Juno

It is time again for our PTL post-Summit webinar series. Come listen to the latest project updates from Icehouse to Juno. Each webinar includes 45 minutes of updates with 15 minutes of q&a. These webinars were established to reduce the number of conflicts during the Summit and allow for broader pa

Re: [Openstack] Renewing keystone signing certs

Hi Adam, Thanks for the advice, I’ve tested it out and it is possible to switch over pretty seamlessly. Here is what I did (spelt out in full for others reading): 1. Generate a new signing key 2. Generate a new certificate request 3. Sign this with the existing CA to generate a new signing_cert

[Openstack] [Swift] Allowing clients to write to separate regions

Hi, I am trying to understand the notion of "regions" in Swift. To start with, it's kinda confusing that the notion of "region" in Keystone is not exactly the same as that of Swift. So I could authenticate with Keystone, get a Swift endpoint for a region (Keystone's notion of a region) and write/r

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

There is an expectation here. Is it explicitly optioned in the API? Should it be? Should a suspended instance be immediately resumable? If the expectation is that a suspended instance is resumable, then the claim against quota should be preserved, and the current behavior is correct. If the expe

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

Playing the cynic, in a cloud ecosystem, users are "greedy". If there is a benefit to the organization of suspending, then an ideal situation would pass along a benefit to the user as well. Otherwise, the busy user is disinclined to suspend because doing so has a cost in the form of instance stat

Re: [Openstack] [Swift] Allowing clients to write to separate regions

Do you plan to have two keystone servers in each region or single keystone server for both east/west coast Swift proxy? 1. Geo-DNS + single Swift region endpoint in keystone 2. Geo-DNS for Keystone servers and each Keystone server returns the local Swift endpoint. 3. Let user to switch which regio

Re: [Openstack] Why doesn't suspend release vCPUs/memory?

I think if we do this, a serious security risk is imported, think this use case: 1) an user has quotas like 10 instances, 20 vcpus, 100G ram and 200G disks 2) he boots 10 instances under his quotas 3) he suspends all this instances 4) he repeats step 2&3 day and night 5) then the cloud platform wil

Re: [Openstack] [Swift] Allowing clients to write to separate regions

I don't plan to use Keystone at all. On Mon, Jun 23, 2014 at 8:13 PM, Kuo Hugo wrote: > Do you plan to have two keystone servers in each region or single keystone > server for both east/west coast Swift proxy? > > 1. Geo-DNS + single Swift region endpoint in keystone > 2. Geo-DNS for Keystone ser

Re: [Openstack] [Swift] Allowing clients to write to separate regions

Got it. So what's the authentication method would you plan to have on Swift now? Basically, you can use Geo-DNS with a single Swift endpoint CNAME record on auth endpoint. Once the auth request be routed to nearest Swift Proxy, it returns it's storage url for the user. 2014-06-24 11:29 GMT+08:

Re: [Openstack] [Swift] Allowing clients to write to separate regions

One more thing, do you read: https://swiftstack.com/blog/2012/09/16/globally-distributed-openstack-swift-cluster/ On Mon, Jun 23, 2014 at 9:57 PM, Michael Gale wrote: > Hello, > >How are you planning to replicate data between regions? You said you > don't want container-sync. > > Also Swift

Re: [Openstack] [Swift] Allowing clients to write to separate regions

Hello, How are you planning to replicate data between regions? You said you don't want container-sync. Also Swift offers read affinity and write affinity, I believe this is setup on the Swift proxy. The affinity settings allow the proxy servers to restrict read and write requests to local reso