[Openstack] [Nova]Update glance image contents

[Ondrej Vaško]

Hello guys,

I am dealing with one issue and that is a question *What is the right
approach for updating openstack glance images?*

When new version of cloud image comes out, for example Ubuntu cloud images
, I want to update old
openstack glance image with that new cloud image (As currently urged with
Spectre mitigation update). I see several possibilities:

   1. Create new image with different name.
>
>Downside is that I will have many images of same distribution and
>release with different updates/kernel and tenants wouldn't know which to
>use. Also there will be additional disk space used.
>
   2. Create new image with same name as old image and delete the old one
   (I think a lot of people is doing it like that).

>Downside is that instances which used the old image will have empty
>image name in horizon and rebuild or other operations may not be working.
>Basically the `*image_ref`* column in nova.instances is pointing to
>UUID of old image and all operations with this UUID will potentially fail.
>
   3. Create new image with different name but hide old image by making it
   private and with no members.

>Hacky way and other issues us above.
>
   4. Create new image, delete the old one and change column `*image_ref*`
   in nova database table instances using the old one to new one.

>Hacky way and possible negative impact.
>

Also notice that API function for glance upload

is working only for images with status *queued* and cannot be used.

So is there a preferred way to update glance image to new or is this still
an open and not addressed issue?

I found similar openstack mailing list conversation in this link
, but no systematic answer
was there so I am asking again here after 3 years.

Thank you forward for your shared insight, advices and experiences.
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

Re: [Openstack] [Nova]Update glance image contents

[Belmiro Moreira]

Hi Ondrej,
the following spec tries to address the issue that you described.
https://review.openstack.org/#/c/508133/

Let me know if you have comments/suggestions.

cheers,
Belmiro

On Fri, Jan 12, 2018 at 2:31 PM, Ondrej Vaško 
wrote:

> Hello guys,
>
> I am dealing with one issue and that is a question *What is the right
> approach for updating openstack glance images?*
>
> When new version of cloud image comes out, for example Ubuntu cloud images
> , I want to update old
> openstack glance image with that new cloud image (As currently urged with
> Spectre mitigation update). I see several possibilities:
>
>1. Create new image with different name.
>>
>>Downside is that I will have many images of same distribution and
>>release with different updates/kernel and tenants wouldn't know which to
>>use. Also there will be additional disk space used.
>>
>2. Create new image with same name as old image and delete the old one
>(I think a lot of people is doing it like that).
>
>>Downside is that instances which used the old image will have empty
>>image name in horizon and rebuild or other operations may not be working.
>>Basically the `*image_ref`* column in nova.instances is pointing to
>>UUID of old image and all operations with this UUID will potentially fail.
>>
>3. Create new image with different name but hide old image by making
>it private and with no members.
>
>>Hacky way and other issues us above.
>>
>4. Create new image, delete the old one and change column `*image_ref*`
>in nova database table instances using the old one to new one.
>
>>Hacky way and possible negative impact.
>>
>
> Also notice that API function for glance upload
> 
> is working only for images with status *queued* and cannot be used.
>
> So is there a preferred way to update glance image to new or is this still
> an open and not addressed issue?
>
> I found similar openstack mailing list conversation in this link
> , but no systematic
> answer was there so I am asking again here after 3 years.
>
> Thank you forward for your shared insight, advices and experiences.
>
> ___
> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/
> openstack
> Post to : openstack@lists.openstack.org
> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/
> openstack
>
>
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

[Openstack] Cinder iscsi performance

[Jimmy Colestock]

Hello All, 

Fighting an issue with cinder iscsi performance being much slow than local 
ephemeral disk.  We’re running multi-path with dual 10Gb nic, offloading is 
working and bandwidth across the links is minimal.   I’ve looked at kernel 
schedulers in the VM’s and tried them all.  

I still need to dig into the storage side (3Par), but just curious if anyone 
has any suggestions while I keep troubleshooting.  

Thanks for any help.

JC
 

Jim Colestock
jcolest...@gmail.com 
https://www.linkedin.com/in/jcolestock/ 

___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

Re: [Openstack] Cinder iscsi performance

[Ivan Kolodyazhny]

Hi Jim,

What target driver do you use? LIO should be faster than a default tgtd.

Regards,
Ivan Kolodyazhny,
http://blog.e0ne.info/

On Fri, Jan 12, 2018 at 4:27 PM, Jimmy Colestock 
wrote:

> Hello All,
>
> Fighting an issue with cinder iscsi performance being much slow than local
> ephemeral disk.  We’re running multi-path with dual 10Gb nic, offloading is
> working and bandwidth across the links is minimal.   I’ve looked at kernel
> schedulers in the VM’s and tried them all.
>
> I still need to dig into the storage side (3Par), but just curious if
> anyone has any suggestions while I keep troubleshooting.
>
> Thanks for any help.
>
> JC
>
>
> Jim Colestock
> jcolest...@gmail.com
> https://www.linkedin.com/in/jcolestock/
>
>
> ___
> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/
> openstack
> Post to : openstack@lists.openstack.org
> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/
> openstack
>
>
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

[Openstack] [Gnocchi][Ceilometer]Alter storage time in gnocchi

[pablo brunetti]

Hello guys,
I installed for my project the Openstack Pike Multinode with 1 controller node 
and 2 controller compute.


I need update the time storage in database metrics for one minute or less. For 
default the gnocchi store each metric in 5 minutes. I upated in 
[dispatcher_gnocchi] for policy medium or high, but the store has not changed. 
The logs archive are ok.


I found similar issue  in ask openstack:
https://ask.openstack.org/en/question/111236/how-to-change-telemetry-measurement-interval-with-gnocchi/
but no systematic answer was there.

Thanks.






___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack