Ask Your Question
0

Copying Image on compute node to avoid glance copying time

asked 2016-01-16 09:30:21 -0500

Awais gravatar image

According to my understanding the Compute node receives the Image from glance and then that image is stored in the cache and served as base to make other instances from it (stored in "/var/lib/nova/instances/_base") directory. Now if i want to overcome the overhead of downloading the image from glance and place the image on compute node (like pre-configured compute node) to start making VMs using that image, (because in my case the image size is really big and i don't need an overhead due to slow internet to make things slower). Would that serve the purpose. I don't want the glance to download the image as that would take time. so i want to place the image on compute server to overcome that delay. Can anybody suggest any optimizations or way to achieve this?

edit retag flag offensive close merge delete

Comments

Did you find out any solution about this?

pamsoo gravatar imagepamsoo ( 2018-03-19 09:23:00 -0500 )edit

2 answers

Sort by ยป oldest newest most voted
0

answered 2016-11-22 04:20:41 -0500

We have a similar setup where we use jenkins to bring up an instance using the image overnight. This is working quite well but we know want to find out a way where we can see which images are already copied over to the compute node..

edit flag offensive delete link more
0

answered 2016-01-17 07:34:18 -0500

Could you please elaborate on why the image needs to be that big? Is it an image-based bootable volume? Is it a snapshot of one? Or a generic image with preallocated blocks?

My point is that maybe it would be better to solve the root cause. For example, each time a user creates a snapshot of the instance, it will be uploaded back to Glance registry; an action that will probably time out along with congesting your network.

edit flag offensive delete link more

Comments

I am trying to deploy a VNF that requires a lot of resources and the image size is quite large (around 5 GB). Now if send the compute node with pre-configured image and stuff it would be good, because that should not cause my end customer's network to congest. compute node should be tested and ready

Awais gravatar imageAwais ( 2016-01-18 06:27:13 -0500 )edit

Hey man, could you please explain how was the solution that you implemented?

Alex Sousa gravatar imageAlex Sousa ( 2018-10-26 07:52:59 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

3 followers

Stats

Asked: 2016-01-16 09:30:21 -0500

Seen: 577 times

Last updated: Nov 22 '16