I launched a F17 instance in our Eucalyptus system.
It works. (Awesome!)
I have not done anything to try and break it. Let me know if I should. :)
On Thu, 27 Sep 2012 15:58:24 -0400 Matthew Miller mattdm@fedoraproject.org wrote:
I launched a F17 instance in our Eucalyptus system.
It works. (Awesome!)
:)
I also have our openstack instance up and seemingly working. Let me know if you want an account there.
I have not done anything to try and break it. Let me know if I should. :)
Feel free. :)
we are kind of in a 'lets see how this works before we start setting it up for production use' stage.
Drop by #fedora-admin or #fedora-noc on irc for more realtime discussion.
kevin
On Thu, Sep 27, 2012 at 02:06:01PM -0600, Kevin Fenzi wrote:
I also have our openstack instance up and seemingly working. Let me know if you want an account there.
Yes please.
Drop by #fedora-admin or #fedora-noc on irc for more realtime discussion.
I will shortly. Thanks.
On Thu, 27 Sep 2012, Matthew Miller wrote:
I launched a F17 instance in our Eucalyptus system.
It works. (Awesome!)
I have not done anything to try and break it. Let me know if I should. :)
break away, if you'd like.
Here's what I've been doing to break the clouds.
1. running ftbfs on euca with 8 (now 5 :() instances - this succeeded in exercising libvirt enough to break it on one NC - I have assurances from euca folks that this is being addressed in 3.1.2 due out 'soon'?
2. runnning ftbfs on openstack with 8 instances - this succeeded in showing up really bad disk slowness on openstack. Compiling packages was less than half the speed as on euca for the same pile of pkgs. I let it run for 3 days and it had done less than half the pkgs that euca had compiled in the same time.
3. running tiobench and bonnie++ on the ephemeral and volume-backed space on openstack to find out what's going on and if it is gluster-backing the os or _something else_ going on
-sv
On 2012-10-04 8:17, Seth Vidal wrote:
- running ftbfs on euca with 8 (now 5 :() instances
- this succeeded in exercising libvirt enough to break it on one NC
- I have assurances from euca folks that this is being addressed in
3.1.2 due out 'soon'?
"Soon" == "now". Hope that helps. ;-)
On Sun, 7 Oct 2012, Garrett Holmstrom wrote:
On 2012-10-04 8:17, Seth Vidal wrote:
- running ftbfs on euca with 8 (now 5 :() instances
- this succeeded in exercising libvirt enough to break it on one NC
- I have assurances from euca folks that this is being addressed in
3.1.2 due out 'soon'?
"Soon" == "now". Hope that helps. ;-)
Any ideas how well things will fare if I apply the updates to a live cloud and kick the daemons? It's one piece I'm still a bit fuzzy on - how well euca surives daemon restart for software upgrades.
-sv
On Mon, Oct 8, 2012 at 8:08 AM, Seth Vidal skvidal@fedoraproject.org wrote:
Any ideas how well things will fare if I apply the updates to a live cloud and kick the daemons? It's one piece I'm still a bit fuzzy on - how well euca surives daemon restart for software upgrades.
Today? Poorly. They currently recommend updating the whole cloud at once. On the bright side, I'm hearing enough noise about that that I expect it to change relatively soon.
On Mon, 8 Oct 2012, Garrett Holmstrom wrote:
On Mon, Oct 8, 2012 at 8:08 AM, Seth Vidal skvidal@fedoraproject.org wrote:
Any ideas how well things will fare if I apply the updates to a live cloud and kick the daemons? It's one piece I'm still a bit fuzzy on - how well euca surives daemon restart for software upgrades.
Today? Poorly. They currently recommend updating the whole cloud at once. On the bright side, I'm hearing enough noise about that that I expect it to change relatively soon.
Well I more meant: If I upgrade to 3.1.2 via rpms and then restart the daemons on all systems - am I likely to lose everything I have running or is it just likely to go sideways for a bit?
-sv
On 2012-10-08 14:17, Seth Vidal wrote:
On Mon, 8 Oct 2012, Garrett Holmstrom wrote:
On Mon, Oct 8, 2012 at 8:08 AM, Seth Vidal skvidal@fedoraproject.org wrote:
Any ideas how well things will fare if I apply the updates to a live cloud and kick the daemons? It's one piece I'm still a bit fuzzy on - how well euca surives daemon restart for software upgrades.
Today? Poorly. They currently recommend updating the whole cloud at once. On the bright side, I'm hearing enough noise about that that I expect it to change relatively soon.
Well I more meant: If I upgrade to 3.1.2 via rpms and then restart the daemons on all systems - am I likely to lose everything I have running or is it just likely to go sideways for a bit?
IIRC, nasty things happen if you upgrade to 3.1.2 while instances are running. Other than that, I'm not aware of anything, but then again I'm not on Eucalyptus's server team.
On Wed, 10 Oct 2012, Garrett Holmstrom wrote:
On 2012-10-08 14:17, Seth Vidal wrote:
On Mon, 8 Oct 2012, Garrett Holmstrom wrote:
On Mon, Oct 8, 2012 at 8:08 AM, Seth Vidal skvidal@fedoraproject.org wrote:
Any ideas how well things will fare if I apply the updates to a live cloud and kick the daemons? It's one piece I'm still a bit fuzzy on - how well euca surives daemon restart for software upgrades.
Today? Poorly. They currently recommend updating the whole cloud at once. On the bright side, I'm hearing enough noise about that that I expect it to change relatively soon.
Well I more meant: If I upgrade to 3.1.2 via rpms and then restart the daemons on all systems - am I likely to lose everything I have running or is it just likely to go sideways for a bit?
IIRC, nasty things happen if you upgrade to 3.1.2 while instances are running.
That is not the most heartening of news. Thanks.
Other than that, I'm not aware of anything, but then again I'm not on Eucalyptus's server team.
Any idea who I should ping to find out what horrible things will befall us?
-sv
infrastructure@lists.fedoraproject.org