Can't create a volume with more than 3 GB using Openstack profile

[Roberto Bandini]

Hi,

I made an experiment based on "Openstack" profile.
I've access the web dashboard and tried to create some volumes.
If I stay under 3 GB all is ok, but if I try to create a volume with more than 3 GB I get an error state.

Is there some limit that I can change? 

Thank you,

Roberto

[David M. Johnson]

The profile setup scripts attempt to create LVM PVs and LVs atop unused
physical disks for OpenStack to use. But, not all Cloudlab clusters
have machines with multiple disks, and one cluster doesn't even have
spare partitions on its single disk. Are you running your experiment at
Cloudlab Utah? Those machines only have one disk and it is fully
partitioned and mounted, so the profile setup scripts cannot create PVs
atop real block devices. The scripts thus fall back to creating a file
in the / fs as a backing store for the PV. In this case, the scripts
don't create a large PV; this is likely your problem.

If you need to use Cloudlab Utah or the existing experiment you've
swapped in, you could extend the backing store file and pvresize the PV
atop it, perhaps. I've never tried that, though. It would be simpler
to use a different cluster with machines that have multiple disks.

> Roberto

David

[Roberto Bandini]

Hi David and thank you very much!

Ok, yes, until I've always used Utah's cloud.
I'm trying right know a new experiment using Clemson's cloud (Is it ok?), but while doing login to the dashboard I'm getting "An error occurred authenticating. Please try again later." even I've received the two emails... what's the problem?

Thank you again,

Roberto

[David M. Johnson]

On 12/2/15 9:49 AM, Roberto Bandini wrote:
> Hi David and thank you very much!
>
> Ok, yes, until I've always used Utah's cloud.
> I'm trying right know a new experiment using Clemson's cloud (Is it
> ok?), but while doing login to the dashboard I'm getting "An error
> occurred authenticating. Please try again later." even I've received the
> two emails... what's the problem?

Yes, it's fine to use Clemson or any of the other clusters. I found
your experiment and successfully logged in to the dashboard using the
random password in the "Profile Instructions" panel on the experiment
status page. You could try ctrl-shift-reload in your browser to ensure
there's no cached state in the dashboard page, or just browse to the
Dashboard link in the profile instructions again... we've seen some
weird things with the Dashboard if you try to login before setup
finishes. Maybe you just tried the wrong password?

> Roberto

David

[Roberto Bandini]

Hi David,

maybe it was a cached state problem as you said.
I try to reload without cache and now it's ok.

I tried also to create a volume greater than 3 GB and that's ok too now, using Clemson.

Thank you very much,

Roberto

[David M. Johnson]

With the new version of the OpenStack profile, the LVM issues are fixed.
We've rebuilt the ARM images with the latest partition layout, which is
similar to our x86 machines, and leaves a large amount of the disk
unpartitioned. The OpenStack scripts now attempt to use this free space
as an LVM PV... so now you can create large volumes on Cloudlab Utah.

David

On 12/2/15 4:51 PM, Roberto Bandini wrote:
> Sembra che su clemson il bootstrap sia finalmente andato in porto! Della
> serie "ora capisco perché Utah è sempre libero"!?
>
>
> 2015-12-02 18:50 GMT+01:00 federica....@unifi.it
> <mailto:federica....@unifi.it> <federica....@unifi.it
> <mailto:federica....@unifi.it>>:
>
> Buono a sapersi.
>
> In realtà ero andata avanti di poco.
> Seguendo le istruzioni di alcuni post
>
> <https://udaraliyanage.wordpress.com/2014/05/23/openstack-increase-volume-capacity/comment-page-1/>https://udaraliyanage.wordpress.com/2014/05/23/openstack-increase-volume-capacity/comment-page-1/____
>
> <https://ask.openstack.org/en/question/25519/failed-when-try-to-create-a-volume-from-a-image-using-command-cinder-create-image-id-123-display-name-boot_vol1-3/>https://ask.openstack.org/en/question/25519/failed-when-try-to-create-a-volume-from-a-image-using-command-cinder-create-image-id-123-display-name-boot_vol1-3/____
>
> poi____
>
> # service cinder-scheduler restart____
>
> # service cinder-api restart__ riesco a creare volumi di 10GB però
> poi ho un errore nel passo successivo, cioè fare l'attach al
> cloudify manager. In generale anche da dashboard quando cerco di
> associare un volume a un'istanza fallisce vediamo se con clemson si
> risolve Fedrica __

> --
>
>
> Federica Paganelli, PhD
> CNIT Researcher
> Research Unit at the University of Florence
> c/o Department of Information Engineering
> via S. Marta 3
> 50139 Firenze
> tel. +39 055 2758524 <tel:%2B39%20055%202758524>
> cell. +39 349 8843383 <tel:%2B39%20349%208843383>
> email: federica....@cnit.it <mailto:federica....@cnit.it>, federica....@unifi.it <mailto:federica....@unifi.it>
> webpage: http://lrst.dinfo.unifi.it/people/Paganelli/index.html
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "cloudlab-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to cloudlab-user...@googlegroups.com
> <mailto:cloudlab-user...@googlegroups.com>.
> To post to this group, send email to cloudla...@googlegroups.com
> <mailto:cloudla...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/cloudlab-users/CANqGBmk0JQxdkRv%3DgbtUcP9P0VHPPFLHxyvhfHOevFZjON3BCQ%40mail.gmail.com
> <https://groups.google.com/d/msgid/cloudlab-users/CANqGBmk0JQxdkRv%3DgbtUcP9P0VHPPFLHxyvhfHOevFZjON3BCQ%40mail.gmail.com?utm_medium=email&utm_source=footer>.
> For more options, visit https://groups.google.com/d/optout.

[Roberto Bandini]

Hi David,

I think that's great, we will try soon again the Cloudify bootstrap process also with Utah cluster.

Thank you very much!

Roberto