Re: [Openstack] How to hit a fast case of nested virtualization

[laclasse]

Hi Mike,

See Rhys post on that: http://www.rdoxenham.com/?p=275

HTH


On Sat, Jun 14, 2014 at 6:01 PM, Mike Spreitzer  wrote:

> Suppose I have an OpenStack undercloud, deployed on some Intel Xeon
> E5-2670 boxes running Ubuntu 12.04.  In /proc/cpuinfo I see the following
> flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat
> pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb
> rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology
> nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2
> ssse3 cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic popcnt
> tsc_deadline_timer aes xsave avx lahf_lm ida arat epb xsaveopt pln pts
> dtherm tpr_shadow vnmi flexpriority ept vpid.
>
> Suppose I create a VM in that undercloud, and use DevStack to install
> OpenStack overcloud inside that VM.  Suppose I create some VMs in that
> overcloud.  Now I am doing nested virtualization.  I am told the speed with
> which the overcloud VMs run can vary greatly, depending on certain
> configuration details.  What do I need to do to get my overcloud to create
> fast VMs?
>
> Following are a couple of remarks I got from one source; do they look
> right?
>
>
>
>
>
>
>
>
>
> *For Ubuntu host, you may try add/replace this line in /etc/modules if it
> is an Intel box: options kvm-intel nested=1 If the host machine is a AMD
> box, you don't have to specify the 'nested=1' parameter. The second option
> for nova.conf can be added directly to the [DEFAULT] section, if it is not
> there.  This is also meant for the 'physical' host. libvirt_cpu_mode =
> host-passthrough*
>
>
> Thanks,
> Mike
> ___
> 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

Re: [Openstack] How to recover my vms after devstack was unstacked?

[laclasse]

Those VMs should be able to boot with on a Linux installation with a KVM
hypervisor with no issues, provided that you have the images (or can grab
them from Glance) and can create a .xml definition file for each.
For the xml definitions, you can cheat a little by creating a blank new KVM
VM using tools such as virt-manager and copy the resulting .xml configs or
use virsh to dump an existing one and editing details in them to point
where your VMs root fs and other resources are located ...
See more related information here:
https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization_Administration_Guide/form-Virtualization-Managing_guests_with_virsh-Creating_a_virtual_machine_XML_dump_configuration_file.html



On Wed, Apr 23, 2014 at 5:04 PM, Michael Monette  wrote:

> Hey all
>
> I know I will get the usual "don't use dev stack in prod" lecture.
> Hopefully we can skip that part. I know now.
>
> We had setup openstack just as a cool project using dev stack. We had to
> move offices and now realized there is some good work in some of the VMS.
>
> I'm locked out of horizon right now and I don't really care about making
> openstack work. I just need to bring up the vms to pul off some data.
> (postgres database). Then I can get rid of all the vms
>
> Is there a way to launch these vms without openstack dashboard? Could I
> use virt manager to start these bad boys?
>
> I have full root on the system. I'm tired of fighting with horizon now I
> just want the vms. Hopefully something can be done.
>
> Yes I know dev stack shouldn't be run in prod :p
>
> Mike
> ___
> 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

Re: [Openstack] Change parameters of kvm in openstack

[laclasse]

However, if removing the '--device' line boots the machine successfully,
you might want to simply check if your init ramfs (of the guest) has a
virtio block driver module included (doesn't seems so), include it (might
as well include the virtio network module) and rebuild it on your SuSE
image.

See this:
http://doc.opensuse.org/products/draft/SLES/SLES-admin_sd_draft/cha.boot.html#sec.boot.initrd






On Wed, Oct 9, 2013 at 12:00 PM, Jeff Cai  wrote:

> Hi,
>
> ** **
>
> When I boot a suse linux instance, it reports that “’system’ volume group
> not found”, I found that if I remove the ‘-device’ line in kvm command
> line, it can start successfully.
>
> Could you tell me how to change the parameters of ‘kvm’ in openstack?
>
> ** **
>
> Following is the command line that starts my instance:
>
> ** **
>
> ** **
>
> /usr/bin/kvm ….
>
> -drive
> file=/var/lib/nova/instances/582db50e-56cf-4541-8078-4329b8578d71/disk,if=none,id=drive-virtio-disk0,format=qcow2,cache=none
> 
>
> -device
> virtio-blk-pci,scsi=off,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1
> 
>
> …
>
> ** **
>
> Best Regards
>
> * *
>
> *Jeff Cai*
>
> *
>
> *
>
> ** **
>
> ___
> 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

Re: [Openstack] openstack meta-data

[laclasse]

Hi Gerard,

I am being told that there is now an iterator in OpenStack when you launch
multiple machines with --num-instances but have not verified this myself as
of yet. I will inquire and test further. I fully agree with your comments
do not think you are doing 'it' wrong at all, just a point below:

The hostname of the machine ideally should be related to the resolved
public/floating ipv4 as you point out, but the unique hostname is also
essential to a machine that might not be exposed externally on that
specific network while still being part of intra-instances communications
on the internal/private 10.X.X.X like network (think multi-tiered
applications for example where a database machine is a datastore for
another and does not need to be exposed publicly. Best practices in IaaS
and OpenStack are to use the internal/private network for this). Maybe this
could maybe be done with multiple 'views' in DNS, and match/map all of them
(public/floating ip space, internal/private ip space and hostnames)?


Cheers

Boris


On Mon, Sep 16, 2013 at 9:18 PM, Gerard Bernabeu  wrote:

>  Hi,
>
> we are using cloud-init to do the VM host contextualization (ie: set the
> right hostname). For this to work we need proper OpenStack metadata, for
> instance we'd like the public-hostname to default to the DNS resolved
> public-ipv4 (or ipv6, if present).
>
> Right now the default behaviour is that public-hostname is the same as the
> hostname and the local-hostname, which is the user-defined instance name.
> I'm not sure if current behaviour is a bug or a feature, but I'd expect
> public-hostname to be related to public-ipv4.
>
> Maybe we're missing something obvious or doing something wrong, are we? If
> not, is there any way we can setup OpenStack (NOVA?) to setup the
> public-hostname metadata to the resolved public-ipv4 name automatically for
> each VM?
>
> Thanks,
>
> Gerard Bernabeu
> FermiCloud and FermiGrid Services at Fermilab
> Phone (+1) 630-840-6509
>
> On 09/15/2013 07:54 PM, laclasse wrote:
>
>  Agreed, unique FQDNs on every instances is still a big issue for common
> workloads, hopefully solved by DNS services coming to OpenStack soon... I
> fully agree that the instance name and the FQDN of the instances should be
> 2 separate things, and even better, I think instances by default should try
> a reverse lookup first when they get on the network, and if available,
> collect a hostname, set it locally and display it in an additional column
> in the DB/tables/webUI, and allows OpenStack services clients to use either
> instance ID, name or hostname.
>
>  What the OP is trying to do is also feasible by installing a bind
> service somewhere on the same subnet (bind9 minimum) and using the
> $GENERATE directive in your zone file, see here on how to write your zone
> file using that directive:
> http://www.oit.uci.edu/dcslib/bind/bind-9.2.1/Bv9ARM.ch06.html#AEN4097
>
>  FWIW, AWS EC2 has a simple iterator that composes a unique hostname to
> any instance by parsing the public IP address, replacing '.' (dots) with
> '-' (to make it viable in a long hostname) and adding prefix/suffix to
> compose the hostname (compute zones and domain) with added dots for
> example: "ec2-54-211-89-105.compute-1.amazonaws.com" is an instance with
> IP 54.211.89.105 ... it a simple and useful solution. You can check it
> yourself:
>
>  $ host 54.211.89.105
>  105.89.211.54.in-addr.arpa domain name pointer
> ec2-54-211-89-105.compute-1.amazonaws.com.
>
>  Not having a unique FQDN on cloud instances can break many apps, and DNS
> services in IaaS should not be taken for granted (failure may happen), so
> IMHO, the best approach should be:
> - boot instance, get on the network, do a reverse look up on IP to see if
> a DNS server answers with a FQDN , collect it and set it
>  - if the look up fails, then resort to the iterator running and still
> end up with an unique hostname per instance.
>
>  It seems a lot of public services hide this issue by allowing to launch
> only one instance at the time, but AFAIK, if you launch several instances
> at once using the nova CLI, you will end up with ALL instances having the
> same hostname (which is the instance name you have given to nova CLI) ...
> which breaks so many things it is not funny ... and is generally not a good
> idea or best practices either.
>
>  If this functionality lands (very soon we hope), it will obviously also
> allows users to deal with hostnames directly rather than IP address, which
> in 2013, could give an archaic and old fashioned unfinished impression to
> users.
>
>  Cheers, sorry for the mini-rant but I would think this is one of the
> most urgent issues to address to see real a

Re: [Openstack] Launch a VM instance with an attached cd-rom?

[laclasse]

Hi,

 What are you trying to do? Do you need to attach/detach a cdrom often? If
it is to install the OS, you should either try a vendor shipped image or
create a custom one, save it and push it to your OpenStack deployment.
 See here:
​​
http://docs.openstack.org/trunk/openstack-compute/admin/content/starting-images.html


If it is for other purposes, you could simply use dd to copy the content of
your iso to a block volume from an instance, once done detach that block
volume and re-attach to another instance. To do this you could follow these
steps:
- launch an instance with an ephemeral/local disk large enough to hold your
.iso / cdrom contents (should not be hard)
- make your cdrom
 contents an .iso file: pop your
​CD​
 in a cdrom on a Linu
​x or ​U
nix based machine and do:
  $ dd if=/dev/cdrom of=/path/to/mycdrom-contents.iso  (assuming
/dev/cdrom is a symlink pointing to your cd player device)
- upload the resulting file using scp to your instance:
  $ scp /path/to/
mycdrom-contents.iso
​  -i myseckey.pem @XXX.XXX.XXX.XXX:~
- create a new block volume of the size of your CD contents (a little bit
over will not hurt) and name it something related to your CD content.
- attach the block volume to your instance
​- ssh to your instance and ​
​
​​again use dd to copy the .iso file to the newly attached block device
(_without_ first making a file system on it)

  $ dd if=/home//
mycdrom-contents.iso
​ of=/dev/sdX (where sdX is the newly attached block device)​
​- detach the block device and attach/re-attach to other instances to your
heart content.
 If the second part did not make sense to you, I would recommend looking at
the documentation at ​
​
http://docs.openstack.org/trunk/

Hope this helps.



On Mon, Sep 16, 2013 at 8:19 AM, Nicolae Paladi  wrote:

> Hi,
>
> is it possible to launch an instance with additional media (e.g. cdrom)
> attached?
> I haven't found anything like that among the nova boot options
>
> cheers,
> /nicolae
>
> ___
> 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

Re: [Openstack] openstack instance does not make use of all of disk space of flavor

[laclasse]

Understood, in this case, I would recommend you use Ubuntu 12.04 LTS images
for your testing.
Hope this helps.


On Mon, Sep 16, 2013 at 2:07 AM, sam lee  wrote:

> @laclasse, @Ritesh, thanks for the help. I will try right now.
>
> @laclasse, I am quite a newbe for openstack, and ubuntu 11.10 is just for
> test.
>
>
> 2013/9/16 laclasse 
>
>>  @sam lee, if I understand properly
>> you are talking about a custom Ubuntu image you created? IIRC all Ubuntu
>> provided default images for OpenStack/AWS after 10.04 LTS have this package
>> installed (or was it starting at 12.04 LTS? Scott Moser the maintainer of
>> the packahe might know more).
>>
>>
>> Also, from your side, I would strongly reconsider and question why you
>> are deploying 11.10 Ubuntu, it is not an LTS release (Long Term Support) an
>> it is already End of Life (a.k.a not supported anymore, see here:
>> https://wiki.ubuntu.com/Releases
>> )
>> ;
>>  you will not get
>> any
>> updated software not mentioning potential security issues.
>>
>> A quick tip, the Ubuntu releases numbers have a meaning: 11 -> Year of
>> release, 10 -> Month of release. So Ubuntu 11.10 was released in October
>> 2011, nearly 2 years ago, in the Linux word, it is legacy and the
>> equivalent of deploying an old Windows release. If an ISV or an specific
>> application forces you to do deploy this version, you should simply push
>> back or give further details on the use case to see if the community can
>> help you further.
>>
>> The exact package you need installed in the instance is called "
>> cloud-initramfs-growroot
>> ", here is its description from the '
>> apt-cache show cloud-initramfs-growroot
>> ' command:
>>
>> Package: cloud-initramfs-growroot
>> Priority: extra
>> Section: universe/admin
>> Installed-Size: 48
>> Maintainer: Scott Moser 
>> Architecture: all
>> Source: cloud-initramfs-tools
>> Version: 0.19ubuntu1
>> Depends: cloud-utils (>= 0.21ubuntu1), initramfs-tools, util-linux (>=
>> 2.17.2)
>> Filename:
>> pool/universe/c/cloud-initramfs-tools/cloud-initramfs-growroot_0.19ubuntu1_all.deb
>> Size: 5692
>> MD5sum: 98035f2475531eec3b3179aeaa56a1d5
>> SHA1: 61a69b041ac8b54153ac6d1c4f9995b5f69b0a65
>> SHA256: 4ca1ec553c6a28a6942a13ea6f2c6db9e175449781a009c008191c19684b0d12
>> Description-en: automatically resize the root partition on first boot
>>  This package adds functionality to an initramfs built by initramfs-tools.
>>  When installed, the initramfs will repartition a disk to make the
>>  root volume consume all space that follows it.
>>   .
>>  You most likely do not want this package unless you know what you are
>>  doing.  It is primarily interesting in a virtualized environment when
>>  a disk can provisioned with a size larger than its original size.
>>  In this case, with this package installed, you can automatically use
>>  the new space without requiring a reboot to re-read the partition table.
>> Homepage: http://launchpad.net/cloud-initramfs-tools
>> Description-md5: 2a0d4bed7bada9873cf69d658abe0c23
>> Bugs: https://bugs.launchpad.net/ubuntu/+filebug
>> Origin: Ubuntu
>>
>>
>> Hope this helps.
>>
>>
>>
>>
>>
>> On Sun, Sep 15, 2013 at 8:09 AM, Ritesh  wrote:
>>
>>> Hello sorry the spell check made package name changed its initramfs grow
>>> root deb package.
>>>
>>> Sent from my iPad
>>>
>>> On 15-Sep-2013, at 12:32 PM, Ritesh  wrote:
>>>
>>> > Hello Sam,
>>> >
>>> > You need to install intramuscular-grow root deb available in Ubuntu ,
>>> which grow your root partition as space available.
>>> >
>>> > Cheers
>>> > Rite an
>>> >
>>> > Sent from my iPad
>>> >
>>> > On 15-Sep-2013, at 9:30 AM, sam lee  wrote:
>>> >
>>> >> I have created a new instance with Ubuntu 11.10 with 80G disk space,
>>> but when I log into the instance and execute "df -h" the space show as
>>> attached dfh.png. and the output of "fdisk -l" as fdisk.png.
>>> >>
>>> >> I want vda taking all of the space and  do two steps as below:
>>> >>
>>> >> 1. fdisk /dev/vda, and create a extended partit
>>> >> 2. mkfs.ext4 /dev/vda1. ==> It will report "/dev/vda is is use"
>>> >>
>>> >> Is this right? If not, what is the correct way to taking 

Re: [Openstack] openstack meta-data

[laclasse]

Agreed, unique FQDNs on every instances is still a big issue for common
workloads, hopefully solved by DNS services coming to OpenStack soon... I
fully agree that the instance name and the FQDN of the instances should be
2 separate things, and even better, I think instances by default should try
a reverse lookup first when they get on the network, and if available,
collect a hostname, set it locally and display it in an additional column
in the DB/tables/webUI, and allows OpenStack services clients to use either
instance ID, name or hostname.

What the OP is trying to do is also feasible by installing a bind service
somewhere on the same subnet (bind9 minimum) and using the $GENERATE
directive in your zone file, see here on how to write your zone file using
that directive:
http://www.oit.uci.edu/dcslib/bind/bind-9.2.1/Bv9ARM.ch06.html#AEN4097

FWIW, AWS EC2 has a simple iterator that composes a unique hostname to any
instance by parsing the public IP address, replacing '.' (dots) with '-'
(to make it viable in a long hostname) and adding prefix/suffix to compose
the hostname (compute zones and domain) with added dots for example: "
ec2-54-211-89-105.compute-1.amazonaws.com" is an instance with IP
54.211.89.105 ... it a simple and useful solution. You can check it
yourself:

$ host 54.211.89.105
105.89.211.54.in-addr.arpa domain name pointer
ec2-54-211-89-105.compute-1.amazonaws.com.

Not having a unique FQDN on cloud instances can break many apps, and DNS
services in IaaS should not be taken for granted (failure may happen), so
IMHO, the best approach should be:
- boot instance, get on the network, do a reverse look up on IP to see if a
DNS server answers with a FQDN , collect it and set it
- if the look up fails, then resort to the iterator running and still end
up with an unique hostname per instance.

It seems a lot of public services hide this issue by allowing to launch
only one instance at the time, but AFAIK, if you launch several instances
at once using the nova CLI, you will end up with ALL instances having the
same hostname (which is the instance name you have given to nova CLI) ...
which breaks so many things it is not funny ... and is generally not a good
idea or best practices either.

If this functionality lands (very soon we hope), it will obviously also
allows users to deal with hostnames directly rather than IP address, which
in 2013, could give an archaic and old fashioned unfinished impression to
users.

Cheers, sorry for the mini-rant but I would think this is one of the most
urgent issues to address to see real automation and large scale apps.




On Sun, Sep 15, 2013 at 12:39 PM, Tim Bell  wrote:

>
> I don’t think there is a formal exit inside OpenStack to do this.  You may
> have to make a small patch to Nova.
>
> It would be an interesting feature to define a plug-in architecture for
> this which would have several other uses.
>
> Tim
>
> > -Original Message-
> > From: Francesco Ceccarelli [mailto:cecca...@fnal.gov]
> > Sent: 14 September 2013 20:40
> > To: Tim Bell
> > Cc: openstack@lists.openstack.org
> > Subject: Re: [Openstack] openstack meta-data
> >
> > Hello,
> > thanks for the help! I hadn't seen the command in the documentation,
> this can be helpful.
> > BTW my problem is to update metadata every time a machine is launched.
> > For example let's talk about the hostname. Every time I launch a
> machine, I would like to have well-configured metadata, with the
> > parameter public-hostname already set to a customized value. I don't
> want the instance name as public-hostname.
> >
> > I hope I was clear and thanks again for the help, Francesco
> >
> >
> >
> > On Sat, 2013-09-14 at 07:18 +, Tim Bell wrote:
> > > You can update the metadata associated with an instance using the
> 'nova meta' command. However, this only updates the metadata
> > (i.e.
> > > key value pairs) rather than the other attributes.
> > >
> > > Changing the server name would not alter the hostname of the VM, it
> would only alter the description in Nova.
> > >
> > > Can you explain a little more what you would like to be doing ?
> > >
> > > Tim
> > >
> > > > -Original Message-
> > > > From: Francesco Ceccarelli [mailto:cecca...@fnal.gov]
> > > > Sent: 14 September 2013 00:11
> > > > To: openstack@lists.openstack.org
> > > > Subject: [Openstack] openstack meta-data
> > > >
> > > > Is it possible to update instance meta-data, for example to set the
> public hostname?
> > > >
> > > > Thanks in advance,
> > > > Francesco
> > > >
> > > >
> > > > ___
> > > > 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.opens

Re: [Openstack] Nova Client CLI Tool: Configure nova client CLI tool to consume image APIs from Glance service

[laclasse]

There is some overlap, maybe due to certain public cloud services that have
decided to expose the glance API through nova, rather than exposing the
glance api end point directly. Not entirely sure of  your end goal, but if
this is a private/custom install of OpenStack and you have the glance
service end point exposed, maybe you could use the 'python-glanceclient'
instead to speak the glance API directly?
Hope this helps.


On Sun, Sep 15, 2013 at 7:29 AM, GROSZ, Maty (Maty) <
maty.gr...@alcatel-lucent.com> wrote:

>  I f I understand you correctly – is there a duplication? Should not all
> the API you mentioned be pointed to Glance already?
>
> Why having these APIs under Nova if you have Glance?
>
> ** **
>
> *From:* Iccha Sethi [mailto:iccha.se...@rackspace.com]
> *Sent:* Thursday, September 12, 2013 20:31
> *To:* GROSZ, Maty (Maty)
> *Cc:* openstack@lists.openstack.org
> *Subject:* RE: [Openstack] Nova Client CLI Tool: Configure nova client
> CLI tool to consume image APIs from Glance service
>
> ** **
>
> Hey,
>
>  
>
> if you look at the readme :https://github.com/openstack/python-novaclient*
> ***
>
>  
>
> there are image calls you can make from nova cli
>
>  
>
> eg:
>
> image-create Create a new image by taking a snapshot of a running server.
> image-delete Delete an image. image-list Print a list of available images
> to boot from. image-meta Set or Delete metadata on an image. image-show
> Show details about the given image.
>
> Thanks,
>
> Iccha
>
>  
>
> -Original Message-
> From: "GROSZ, Maty (Maty)" 
> Sent: Thursday, September 12, 2013 4:43am
> To: "openstack@lists.openstack.org" 
> Subject: [Openstack] Nova Client CLI Tool: Configure nova client CLI tool
> to consume image APIs from Glance service
>
> Hey,
>
> I am trying to find out how do I configure nova client CLI  tool to
> consume image APIs from Glance service instead of nova v2.0.
>
> Someone can help?
>
> Thanks,
>
> Maty.
>
> [image: logo]
>
> *Maty Grosz*
>
> Alcatel-Lucent
>
> APIs Functional Owner, R&D
>
> CLOUDBAND BUSINESS UNIT
>
> 16 Atir Yeda St. Kfar-Saba 44643, ISRAEL
>
> T: +972 (0) 9 7933078
>
> F: +972 (0) 9 7933700
>
> *maty.gr...@alcatel-lucent.com * 
>
>  
>
>  
>
> ___
> 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

Re: [Openstack] openstack instance does not make use of all of disk space of flavor

[laclasse]

@sam lee, if I understand properly
​
​ you are talking about a custom Ubuntu image you created? IIRC all Ubuntu
provided default images for OpenStack/AWS after 10.04 LTS have this package
installed (or was it starting at 12.04 LTS? Scott Moser the maintainer of
the packahe might know more).


Also, from your side, I would strongly reconsider and question why you are
deploying 11.10 Ubuntu, it is not an LTS release (Long Term Support) an it
is already End of Life (a.k.a not supported anymore, see here:
https://wiki.ubuntu.com/Releases
​)​
;​
 you will not get
​ any ​
updated software not mentioning potential security issues.

​A quick tip, the Ub​untu releases numbers have a meaning: 11 -> Year of
release, 10 -> Month of release. So Ubuntu 11.10 was released in October
2011, nearly 2 years ago, in the Linux word, it is legacy and the
equivalent of deploying an old Windows release. If an ISV or an specific
application forces you to do deploy this version, you should simply push
back or give further details on the use case to see if the community can
help you further.

The exact package you need installed in the instance is called "​
cloud-initramfs-growroot
​", here is its description from the ​'
apt-cache show cloud-initramfs-growroot
​' command:

Package: cloud-initramfs-growroot
Priority: extra
Section: universe/admin
Installed-Size: 48
Maintainer: Scott Moser 
Architecture: all
Source: cloud-initramfs-tools
Version: 0.19ubuntu1
Depends: cloud-utils (>= 0.21ubuntu1), initramfs-tools, util-linux (>=
2.17.2)
Filename:
pool/universe/c/cloud-initramfs-tools/cloud-initramfs-growroot_0.19ubuntu1_all.deb
Size: 5692
MD5sum: 98035f2475531eec3b3179aeaa56a1d5
SHA1: 61a69b041ac8b54153ac6d1c4f9995b5f69b0a65
SHA256: 4ca1ec553c6a28a6942a13ea6f2c6db9e175449781a009c008191c19684b0d12
Description-en: automatically resize the root partition on first boot
 This package adds functionality to an initramfs built by initramfs-tools.
 When installed, the initramfs will repartition a disk to make the
 root volume consume all space that follows it.
 .
 You most likely do not want this package unless you know what you are
 doing.  It is primarily interesting in a virtualized environment when
 a disk can provisioned with a size larger than its original size.
 In this case, with this package installed, you can automatically use
 the new space without requiring a reboot to re-read the partition table.
Homepage: http://launchpad.net/cloud-initramfs-tools
Description-md5: 2a0d4bed7bada9873cf69d658abe0c23
Bugs: https://bugs.launchpad.net/ubuntu/+filebug
Origin: Ubuntu


Hope this helps.



​


On Sun, Sep 15, 2013 at 8:09 AM, Ritesh  wrote:

> Hello sorry the spell check made package name changed its initramfs grow
> root deb package.
>
> Sent from my iPad
>
> On 15-Sep-2013, at 12:32 PM, Ritesh  wrote:
>
> > Hello Sam,
> >
> > You need to install intramuscular-grow root deb available in Ubuntu ,
> which grow your root partition as space available.
> >
> > Cheers
> > Rite an
> >
> > Sent from my iPad
> >
> > On 15-Sep-2013, at 9:30 AM, sam lee  wrote:
> >
> >> I have created a new instance with Ubuntu 11.10 with 80G disk space,
> but when I log into the instance and execute "df -h" the space show as
> attached dfh.png. and the output of "fdisk -l" as fdisk.png.
> >>
> >> I want vda taking all of the space and  do two steps as below:
> >>
> >> 1. fdisk /dev/vda, and create a extended partit
> >> 2. mkfs.ext4 /dev/vda1. ==> It will report "/dev/vda is is use"
> >>
> >> Is this right? If not, what is the correct way to taking all of the
> space?
> >>
> >> Thanks in advance.
> >>
> >>
> >> 
> >> 
> >> ___
> >> 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
>
___
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] Does openstack python API support vsphere?

[laclasse]

Install python-suds on the nodes to speak to the vsphere API.
See http://docs.openstack.org/trunk/openstack-compute/admin/content/vmware.html

Hope this helps.




On Wed, Sep 4, 2013 at 3:03 AM, openstack learner
 wrote:
> Hi all,
>
> I am new to openstack, can anyone tell me if the openstack nova python API
> support vsphere?
>
> Thanks a lot
>
> xin
>
> ___
> 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

Re: [Openstack] can instances be resized?

[laclasse]

not dynamically no. You can launch new and additional instances of
different flavours across the same cloud (scale out), but "scale in"
the instances dynamically is not supported by default atm.

On Fri, Aug 30, 2013 at 8:46 AM, Mark Chaney  wrote:
> Thanks for that tip and link. Can you only change flavors of an instance or
> can you granularly adjust an intense? If I cant,that means I have to make a
> flavor for every little custom difference to an existing flavor?
>
>
> On 2013-08-30 02:18, laclasse wrote:
>>
>> That is correct, it does.
>> Takes care of extending the rootfs of the instance according to the
>> flavours (sizes of VM) that you are defining/modifying or using the
>> pre-defined ones. See
>> http://docs.openstack.org/trunk/openstack-ops/content/flavors.html
>>
>> On Fri, Aug 30, 2013 at 7:15 AM, Mark Chaney 
>> wrote:
>>>
>>> I am going to be hosting mainly Ubuntu, CentOS, and some Debian instances
>>> with my openstack "cloud". Can instances be resized when it comes to
>>> their
>>> storage? Obviously not only does its disk need to be resized, but its FS
>>> as
>>> well. I am hoping openstack can take care of all of that.
>>>
>>> Thanks,
>>> Mark
>>>
>>> ___
>>> 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

Re: [Openstack] can instances be resized?

[laclasse]

That is correct, it does.
Takes care of extending the rootfs of the instance according to the
flavours (sizes of VM) that you are defining/modifying or using the
pre-defined ones. See
http://docs.openstack.org/trunk/openstack-ops/content/flavors.html

On Fri, Aug 30, 2013 at 7:15 AM, Mark Chaney  wrote:
> I am going to be hosting mainly Ubuntu, CentOS, and some Debian instances
> with my openstack "cloud". Can instances be resized when it comes to their
> storage? Obviously not only does its disk need to be resized, but its FS as
> well. I am hoping openstack can take care of all of that.
>
> Thanks,
> Mark
>
> ___
> 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

Re: [Openstack] Import Centos 5 (KVM) to Openstack

[laclasse]

There is a good chance that your centos 5.6 image was not made on
virtIO block device for the root disk, hence the error trying to boot
on one.
Include a virtIO block driver in your base init ramdisk and rebuild it
(the ramdisk of the VM). Then it should work.

See:
https://bugzilla.redhat.com/show_bug.cgi?id=560672
http://crunchtools.com/kvm-virtualization-102/
http://wiki.libvirt.org/page/Virtio

Hope this helps.

On Wed, Aug 28, 2013 at 9:24 AM, Mahardhika
 wrote:
> Hi, wondering if you guys have tried to import centos 5.6 that running
> previously on KVM to Openstack infrastructure
> while importing centos 6 it's work well, but not with centos 5.6, it said
> No Volume groups found
> Volume group VolGroup00 not foumd
> kernel pani
> etc
>
> anyone have tried this?
> --
> Regards,
> Mahardhika Gilang
>
>
> ___
> 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