Re: [CentOS] Single source repo, producing packages for CentOS 6 and 7, how?

[Martinx - ジェームズ]

On 9 August 2016 at 15:50, Peter <pe...@pajamian.dhs.org> wrote:

> On 10/08/16 07:31, Martinx - ジェームズ wrote:
> > Guys,
> >
> >  I have a source code here, that produces RPM packages for CentOS 6, that
> > includes an Upstart Job and a ugly and minimal Init Script.
> >
> >  Now, I have it running on CentOS 7, via systemd, looks good!
> >
> >  However, my RPM package for CentOS 7 still includes the useless Upstart
> > and Init Scripts...
> >
> >  So, here is the question:
> >
> >  How are you guys managing this? I mean, I would like to have 1 single
> > source and 1 SPEC file for both CentOS 6 and 7, but no need to systemd
> > service files on 6 / no need to upstart job files on 7.
>
> %if 0%{?rhel} > 6
> # Stuff for CentOS 7 here
> %else
> # Stuff for CentOS 6 here
> %endif
>

Sweet... Thank you guys!
___
CentOS mailing list
CentOS@centos.org
https://lists.centos.org/mailman/listinfo/centos

[CentOS] Single source repo, producing packages for CentOS 6 and 7, how?

[Martinx - ジェームズ]

Guys,

 I have a source code here, that produces RPM packages for CentOS 6, that
includes an Upstart Job and a ugly and minimal Init Script.

 Now, I have it running on CentOS 7, via systemd, looks good!

 However, my RPM package for CentOS 7 still includes the useless Upstart
and Init Scripts...

 So, here is the question:

 How are you guys managing this? I mean, I would like to have 1 single
source and 1 SPEC file for both CentOS 6 and 7, but no need to systemd
service files on 6 / no need to upstart job files on 7.

Cheers!
Thiago
___
CentOS mailing list
CentOS@centos.org
https://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] CentOS-7-x86_64-GenericCloud.raw very unusual format

[Martinx - ジェームズ]

>
> > Why not a compressed qcow2?
> >
> > Like: CentOS-7-x86_64-GenericCloud.qcow2c
> >
> > This way, we can add it as-is to OpenStack Glance, that it will be
> > downloaded by demand, when required and, OpenStack Nova will
> automatically
> > uncompress the image for runtime.
>
> We do publish a qcow2c, the always updated version ( updated each month
> end ) is at :
> http://cloud.centos.org/centos/7/images/CentOS-7-x86_64-GenericCloud.qcow2c
>
> This is a symlink that points to the latest image. At the moment this is
> :
>
> http://cloud.centos.org/centos/7/images/CentOS-7-x86_64-GenericCloud-1602.qcow2c
>
> There will be a 1603 by the end of this week
>
>
> --
> Karanbir Singh
> +44-207-0999389 | http://www.karan.org/ | twitter.com/kbsingh
> GnuPG Key : http://www.karan.org/publickey.asc


Sounds awesome! Thanks for sharing the symlink... I'll use it instead..
Cheers!
___
CentOS mailing list
CentOS@centos.org
https://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] CentOS-7-x86_64-GenericCloud.raw very unusual format

[Martinx - ジェームズ]

On 21 March 2016 at 11:51, Karanbir Singh  wrote:

> On 21/03/16 14:34, Aliaksei Sheshka wrote:
> > What is the point to pack single raw file into the tar.gz and to name it
> raw ?
> > Or it's just a mistake ?
> >
> >
>
> Johnny is looking at the cdn side of thigs, but effectively we will not
> have a .raw - we will have a .raw.tar.gz file ( which will itself only
> have a .raw file compressed inside it )
>
>
> --
> Karanbir Singh
> +44-207-0999389 | http://www.karan.org/ | twitter.com/kbsingh
> GnuPG Key : http://www.karan.org/publickey.asc
>
>
Why not a compressed qcow2?

Like: CentOS-7-x86_64-GenericCloud.qcow2c

This way, we can add it as-is to OpenStack Glance, that it will be
downloaded by demand, when required and, OpenStack Nova will automatically
uncompress the image for runtime.

Cheers!
Thiago
___
CentOS mailing list
CentOS@centos.org
https://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] CentOS 7 regression, can not take a KVM qcow2, convert it and boot it on ESXi...

[Martinx - ジェームズ]

On 4 February 2016 at 06:58, John Hodrien <j.h.hodr...@leeds.ac.uk> wrote:
> On Thu, 4 Feb 2016, Martinx - ジェームズ wrote:
>
>> My last try was, during Packer build, to append "add_drivers+="
>> sd_mod" to /etc/dracut.conf and re-building initramdisk by running
>> "dracut -f". However, it does not boot either (on ESXi)!
>>
>> I can still boot using Rescue mode on VMWare, then, I can see that
>> sd_mod is inside of initramfs as expected (lsinitrd FTW) but, dracut
>> does not load it! I tried "force_drivers+=" sd_mod", doesn't work
>> too...
>>
>> Weird is that, while during Rescue, if I run "dracut -f" again, then,
>> it works! Normal boot proceeds... But this is unacceptable, because I
>> can not ship a half-broken VMDK and tell customers to first boot on
>> Rescue, run "dracut -f" and then, use the system.
>>
>> So, why the heck the "dracut -f" executed when as a KVM guest have no
>> effect for later ESXi usage?
>>
>> What am I missing?
>>
>> I really appreciate any help!
>
>
> Is this possible the hostonly option in dracut causing fun?
>
> "On RHEL-7 the hostonly mode is the default mode. Generic "non-hostonly"
>  images are created, if the dracut-config-generic rpm is installed.  The
>  rescue kernel entry in the bootloader menu is also a generic image."
>
> jh

You're the man!!! Thank you so much!! YAY!!

I need to install the package "dracut-config-generic" during Packer
build, which sets "hostonly=no" and, voialá!  :-D

Thank you again!

Cheers!
Thiago
___
CentOS mailing list
CentOS@centos.org
https://lists.centos.org/mailman/listinfo/centos

[CentOS] CentOS 7 regression, can not take a KVM qcow2, convert it and boot it on ESXi...

[Martinx - ジェームズ]

Guys,

 I have a Packer build procedure, that works like a charm, when with CentOS 6.7.

 Exemplifying:

 1- Packer (with virtio disk) + QEmu + CentOS 6.7 ISO;
 2- Create a RAW image;
 3- Convert the RAW image into QCOW2 for KVM hypervisors (okay);
 4- Convert the RAW image into VMDK for ESXi hypervisors (okay).

 However, when doing the very same procedure, with CentOS 7.2 ISO, it
does not boot on ESXi!

 After research, I realized that it is not auto-loading the sd_mod on ESXi.

 I'm seeing errors like this:

 dracut-initqueue timeout
 dracut-initqueue timeout 

 And it drops to "dracut#" shell, where I can see that there is no
/dev/sda listed on /proc/partitions.

 I think I tried everything!

 My last try was, during Packer build, to append "add_drivers+="
sd_mod" to /etc/dracut.conf and re-building initramdisk by running
"dracut -f". However, it does not boot either (on ESXi)!

 I can still boot using Rescue mode on VMWare, then, I can see that
sd_mod is inside of initramfs as expected (lsinitrd FTW) but, dracut
does not load it! I tried "force_drivers+=" sd_mod", doesn't work
too...

 Weird is that, while during Rescue, if I run "dracut -f" again, then,
it works! Normal boot proceeds... But this is unacceptable, because I
can not ship a half-broken VMDK and tell customers to first boot on
Rescue, run "dracut -f" and then, use the system.

 So, why the heck the "dracut -f" executed when as a KVM guest have no
effect for later ESXi usage?

 What am I missing?

 I really appreciate any help!

Thanks!
Thiago
___
CentOS mailing list
CentOS@centos.org
https://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] CentOS 7 regression, can not take a KVM qcow2, convert it and boot it on ESXi...

[Martinx - ジェームズ]

On 4 February 2016 at 00:58, Gordon Messmer <gordon.mess...@gmail.com> wrote:
> On 02/03/2016 06:33 PM, Martinx - ジェームズ wrote:
>>
>>   What am I missing?
>
>
> Use lsinitrd to compare the two initrds.  Direct each output to two files
> and then use "diff -u" to see the difference.

Mmm... Nice tip! Thanks!  :-D
___
CentOS mailing list
CentOS@centos.org
https://lists.centos.org/mailman/listinfo/centos