Ask Your Question

Launching large instance makes all VM's on compute node basically unresponsive

asked 2013-06-21 07:06:10 -0500

updated 2013-06-21 14:53:40 -0500

smaffulli gravatar image

Our company is running Grizzly release of openstack. We run very large instance footprints 470GB, when one of these instances is spawning on a compute node for the first time it generally makes all other instances on the compute node unresponsive. I'm pretty sure this is due to the massive amount of IO required to transfer and setup the image for use. This behavior however is extremely undesirable because other VM's can be doing important work.

So my question is, can I limit the resources that can be used for this spawning process? Maybe set nice level or something?

edit retag flag offensive close merge delete


I am still in the testing mode, and have just run into this. I would like to hear on this. I can't ssh in, nor use openstack dashboard to any usefulness. To the point where I am finally getting PO'd and rebooting the entire compute node after 20-30 minutes.

OldParrothead gravatar imageOldParrothead ( 2013-07-11 09:59:04 -0500 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2013-07-14 22:41:08 -0500

skyrainman gravatar image

Hey there,

That's simple, as you mentioned, it's IO. I had this problem a lot, especially with SATA drives. Upgrade your disk to SSD, they're cheap now, the investment will pay off.

Note that any time spent trying to investigate the software is a waste of time.

edit flag offensive delete link more

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



Asked: 2013-06-21 07:06:10 -0500

Seen: 278 times

Last updated: Jul 14 '13