This is a development branch thing but it's related to Cobbler's plan moving forward, so it seems best to discuss here.
So ... there's been some good work going so far towards supporting some variants of Windows, though I see it may perhaps be somewhat of an evolutionary dead-end.
The web page for RIS-Linux in particular does not mention Vista or 2008. Further, I am much more interested in supporting Windows virtually than physically -- this should be natural with the increasing interest in virtualization and recent interoperability agreements with Microsoft.
For the devel branch (1.7), I'd rather we refocus our efforts into making sure the experience for Windows installs, virtually, is as good as possible.
The upside of this is that most of the infrastructure is already in place -- we can already do ISO based fullvirt installs ("cobbler image add" with the ISO file residing on NFS) and a next step is to teach it about a virtual floppy drive with the SIF answer file on the drive, so it can be fully scripted.
We already also have the "virt-clone" image type, for being able to take an existing disk image and repeatedly clone that image with koan, keeping the same source image on NFS. (The syntax here is "koan --image=foo --virt", just like with the ISO based installs for virt).
The goal here is to not invest too much effort in supporting dead-end deployment areas and writing code to cater to say, XP vs Vista vs 2003 vs 2008, but handle things generically, with answer files, and images, things we already do. On the plus side, there's also much less work in doing this and no additional dependencies or things to configure and set up.
Rather than physical deployments this encourages deploying Windows on Linux hosts, which makes the Windows machines easier to manage since you can run tools like libvirt and Func on the hosts. See http://www.redhat.com/promo/svvp/
Windows belongs running on virt. In general, for non-Linux OS's, we should also concentrate on virt.
--Michael
On Mon, Mar 23, 2009 at 9:06 AM, Michael DeHaan mdehaan@redhat.com wrote:
This is a development branch thing but it's related to Cobbler's plan moving forward, so it seems best to discuss here.
So ... there's been some good work going so far towards supporting some variants of Windows, though I see it may perhaps be somewhat of an evolutionary dead-end.
The web page for RIS-Linux in particular does not mention Vista or 2008. Further, I am much more interested in supporting Windows virtually than physically -- this should be natural with the increasing interest in virtualization and recent interoperability agreements with Microsoft.
For the devel branch (1.7), I'd rather we refocus our efforts into making sure the experience for Windows installs, virtually, is as good as possible.
The upside of this is that most of the infrastructure is already in place -- we can already do ISO based fullvirt installs ("cobbler image add" with the ISO file residing on NFS) and a next step is to teach it about a virtual floppy drive with the SIF answer file on the drive, so it can be fully scripted.
We already also have the "virt-clone" image type, for being able to take an existing disk image and repeatedly clone that image with koan, keeping the same source image on NFS. (The syntax here is "koan --image=foo --virt", just like with the ISO based installs for virt).
The goal here is to not invest too much effort in supporting dead-end deployment areas and writing code to cater to say, XP vs Vista vs 2003 vs 2008, but handle things generically, with answer files, and images, things we already do. On the plus side, there's also much less work in doing this and no additional dependencies or things to configure and set up.
Rather than physical deployments this encourages deploying Windows on Linux hosts, which makes the Windows machines easier to manage since you can run tools like libvirt and Func on the hosts. See http://www.redhat.com/promo/svvp/
Windows belongs running on virt.
To play devil's advocate (I am not in this situation) what about the poor people who just want to make deploying windows on physical hosts easier? It seems like this direction hurts those. Why don't we do a dual pronged approach? Make the virtualized windows experience smooth as butter but continue using ris-linux for bare metal windows? Just my 2 cents.
Jeff Schroeder wrote:
On Mon, Mar 23, 2009 at 9:06 AM, Michael DeHaan mdehaan@redhat.com wrote:
This is a development branch thing but it's related to Cobbler's plan moving forward, so it seems best to discuss here.
So ... there's been some good work going so far towards supporting some variants of Windows, though I see it may perhaps be somewhat of an evolutionary dead-end.
The web page for RIS-Linux in particular does not mention Vista or 2008. Further, I am much more interested in supporting Windows virtually than physically -- this should be natural with the increasing interest in virtualization and recent interoperability agreements with Microsoft.
For the devel branch (1.7), I'd rather we refocus our efforts into making sure the experience for Windows installs, virtually, is as good as possible.
The upside of this is that most of the infrastructure is already in place -- we can already do ISO based fullvirt installs ("cobbler image add" with the ISO file residing on NFS) and a next step is to teach it about a virtual floppy drive with the SIF answer file on the drive, so it can be fully scripted.
We already also have the "virt-clone" image type, for being able to take an existing disk image and repeatedly clone that image with koan, keeping the same source image on NFS. (The syntax here is "koan --image=foo --virt", just like with the ISO based installs for virt).
The goal here is to not invest too much effort in supporting dead-end deployment areas and writing code to cater to say, XP vs Vista vs 2003 vs 2008, but handle things generically, with answer files, and images, things we already do. On the plus side, there's also much less work in doing this and no additional dependencies or things to configure and set up.
Rather than physical deployments this encourages deploying Windows on Linux hosts, which makes the Windows machines easier to manage since you can run tools like libvirt and Func on the hosts. See http://www.redhat.com/promo/svvp/
Windows belongs running on virt.
To play devil's advocate (I am not in this situation) what about the poor people who just want to make deploying windows on physical hosts easier? It seems like this direction hurts those.
We don't really have a vested interest in making Windows physical deployments easy to deploy.
Why don't we do a dual pronged approach? Make the virtualized windows experience smooth as butter but continue using ris-linux for bare metal windows? Just my 2 cents.
Ultimately it's about what is the best use of time. Keeping things small and focused in Cobbler is something we must always do, and sometimes this means knowing what we should and shouldn't work on.
This list becoming 1/2 about Windows questions would not be good -- and it distracts us from the core mission.
The virtualization seems ideally suited for dealing with such things, and it keeps them simple -- so we can concentrate on other, perhaps more important, work items. (Like that nice shiny RFE list).
--Michael
Michael DeHaan wrote:
Jeff Schroeder wrote:
On Mon, Mar 23, 2009 at 9:06 AM, Michael DeHaan mdehaan@redhat.com wrote:
This is a development branch thing but it's related to Cobbler's plan moving forward, so it seems best to discuss here.
So ... there's been some good work going so far towards supporting some variants of Windows, though I see it may perhaps be somewhat of an evolutionary dead-end.
The web page for RIS-Linux in particular does not mention Vista or 2008. Further, I am much more interested in supporting Windows virtually than physically -- this should be natural with the increasing interest in virtualization and recent interoperability agreements with Microsoft.
For the devel branch (1.7), I'd rather we refocus our efforts into making sure the experience for Windows installs, virtually, is as good as possible.
The upside of this is that most of the infrastructure is already in place -- we can already do ISO based fullvirt installs ("cobbler image add" with the ISO file residing on NFS) and a next step is to teach it about a virtual floppy drive with the SIF answer file on the drive, so it can be fully scripted.
We already also have the "virt-clone" image type, for being able to take an existing disk image and repeatedly clone that image with koan, keeping the same source image on NFS. (The syntax here is "koan --image=foo --virt", just like with the ISO based installs for virt).
The goal here is to not invest too much effort in supporting dead-end deployment areas and writing code to cater to say, XP vs Vista vs 2003 vs 2008, but handle things generically, with answer files, and images, things we already do. On the plus side, there's also much less work in doing this and no additional dependencies or things to configure and set up.
Rather than physical deployments this encourages deploying Windows on Linux hosts, which makes the Windows machines easier to manage since you can run tools like libvirt and Func on the hosts. See http://www.redhat.com/promo/svvp/
Windows belongs running on virt.
To play devil's advocate (I am not in this situation) what about the poor people who just want to make deploying windows on physical hosts easier? It seems like this direction hurts those.
We don't really have a vested interest in making Windows physical deployments easy to deploy.
Why don't we do a dual pronged approach? Make the virtualized windows experience smooth as butter but continue using ris-linux for bare metal windows? Just my 2 cents.
Ultimately it's about what is the best use of time. Keeping things small and focused in Cobbler is something we must always do, and sometimes this means knowing what we should and shouldn't work on.
This list becoming 1/2 about Windows questions would not be good -- and it distracts us from the core mission.
The virtualization seems ideally suited for dealing with such things, and it keeps them simple -- so we can concentrate on other, perhaps more important, work items. (Like that nice shiny RFE list).
--Michael
I should also point out that I spelled 'strategery' wrong.
Apologies to George :)
Michael DeHaan schrieb:
We don't really have a vested interest in making Windows physical deployments easy to deploy.
;-)
Honi soit qui mal y pense ;-)
This list becoming 1/2 about Windows questions would not be good -- and it distracts us from the core mission.
Indeed. Supporting Windows just sucks too many resources.
Regards, Rainer
On Mon, 23 Mar 2009 17:40:57 +0100, Rainer Duffner rainer@ultra-secure.de wrote:
Michael DeHaan schrieb:
We don't really have a vested interest in making Windows physical deployments easy to deploy.
;-)
Honi soit qui mal y pense ;-)
This list becoming 1/2 about Windows questions would not be good -- and it distracts us from the core mission.
Indeed. Supporting Windows just sucks too many resources.
Regards, Rainer _______________________________________________ cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
http://en.wikipedia.org/wiki/Honi_soit_qui_mal_y_pense, for those who were curious like myself :)
On Mon, Mar 23, 2009 at 16:40, Rainer Duffner rainer@ultra-secure.de wrote:
Michael DeHaan schrieb:
We don't really have a vested interest in making Windows physical deployments easy to deploy.
Hello,
So just to add my 2c in...
This statement worries me, I would suggest while we as a community should encourage the use of FOSS operating systems do we really want to take an RMS style view on this? Cobbler from my POV should be a generic systems management and will work best if it can document and deploy your whole infrastructure, re my earlier comments about upgrading iLO cards and network devices from Cobbler.
Windows is not going to be going away from the data centre any time soon and while I believe everything should be virtual unless there is a good reason not to be not everyone has that view.
This list becoming 1/2 about Windows questions would not be good -- and it distracts us from the core mission.
Indeed. Supporting Windows just sucks too many resources.
Would these support requests not be balanced out by an increased number of Windows users on the list?
Again just my thoughts, if the community wants to provide and support Windows deployment on tin then why not allow it?
Take care.
On Mar 25, 2009, at 8:15 AM, Robert Lazzurs rob@lazzurs.net wrote:
On Mon, Mar 23, 2009 at 16:40, Rainer Duffner <rainer@ultra- secure.de> wrote:
Michael DeHaan schrieb:
We don't really have a vested interest in making Windows physical deployments easy to deploy.
Hello,
So just to add my 2c in...
This statement worries me, I would suggest while we as a community should encourage the use of FOSS operating systems do we really want to take an RMS style view on this? Cobbler from my POV should be a generic systems management and will work best if it can document and deploy your whole infrastructure, re my earlier comments about upgrading iLO cards and network devices from Cobbler.
Windows is not going to be going away from the data centre any time soon and while I believe everything should be virtual unless there is a good reason not to be not everyone has that view.
This is pretty much how I feel about it, but again I'm biased :)
This list becoming 1/2 about Windows questions would not be good -- and it distracts us from the core mission.
Indeed. Supporting Windows just sucks too many resources.
Would these support requests not be balanced out by an increased number of Windows users on the list?
Again just my thoughts, if the community wants to provide and support Windows deployment on tin then why not allow it?
Take care.
My thought regarding this would be to spin off a new list if it ever got that bad.
cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
-- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
Robert Lazzurs wrote:
On Mon, Mar 23, 2009 at 16:40, Rainer Duffner rainer@ultra-secure.de wrote:
Michael DeHaan schrieb:
We don't really have a vested interest in making Windows physical deployments easy to deploy.
Hello,
So just to add my 2c in...
This statement worries me, I would suggest while we as a community should encourage the use of FOSS operating systems do we really want to take an RMS style view on this? Cobbler from my POV should be a generic systems management and will work best if it can document and deploy your whole infrastructure, re my earlier comments about upgrading iLO cards and network devices from Cobbler.
Windows is not going to be going away from the data centre any time soon and while I believe everything should be virtual unless there is a good reason not to be not everyone has that view.
Ok, I'll explain this further.
Cobbler is growing rather fast... VERY fast, and as it grows, I want to make sure the quality of the application everywhere is the best it can be.
There are a lot of things left undone, and various rough corners that still need to be polished. SUSE, Network Config, and Debian/Ubuntu support are all very recent and in varying degrees of being finished. We're not even packaged for all of them. There are still things that can't be done in the web application, etc. Lots of expansion, but we can do more to refine those things. (Does koan work on all of these platforms yet for virt AND replacement? Etc). If we expand too fast, we forget to go back and fill in that detail.
I seem to encounter a lot of users that /do/ clone physical windows systems, so I am not sure doing scripted installs is the answer. Further, the RIS stuff does not seem to be consistent across all OSes . An approach based on cloning those systems may be better in the end because that also allows working systems to be deployed .. not just empty shells that just have the Base OS on them.
I think we need to take a step back for a bit and focus on the core of Cobbler, close out lots of bugs and RFEs, ensure our new distribution support is first-class on the level with our others, and make sure the web application has parity with the command line.
Yes, all of these things /can/ be done in parallel, but as we add new arms to our octopus, we must make sure we go back and add the proverbial suckers to those arms.
Perfectionism is /not/ a goal, but I do want to make sure we don't spread out too fast and leave things unfinished.
Regarding the cloning of physical systems, a while back Andrew Brown contributed a live image that would clone (a bit slowly) a physical system using cobbler to network deploy the clone image. I think it might be interesting to try to resurrect and supercharge that idea. This way we could be sure to deploy Windows systems that /also/ had their software installed, without having to also teach Cobbler about Windows domain automation. We say Clonezilla and other analogs is hard to deploy, why not write a better solution?
And, of course, where virtualization is possible, I think that /IS/ a better route. We should encourage more use of that technology and also enhance our capabilities there where it makes sense.
--Michael
On Wed, Mar 25, 2009 at 10:49 AM, Michael DeHaan mdehaan@redhat.com wrote:
Robert Lazzurs wrote:
On Mon, Mar 23, 2009 at 16:40, Rainer Duffner rainer@ultra-secure.de
wrote:
Michael DeHaan schrieb:
We don't really have a vested interest in making Windows physical deployments easy to deploy.
Hello,
So just to add my 2c in...
This statement worries me, I would suggest while we as a community should encourage the use of FOSS operating systems do we really want to take an RMS style view on this? Cobbler from my POV should be a generic systems management and will work best if it can document and deploy your whole infrastructure, re my earlier comments about upgrading iLO cards and network devices from Cobbler.
Windows is not going to be going away from the data centre any time soon and while I believe everything should be virtual unless there is a good reason not to be not everyone has that view.
Ok, I'll explain this further.
Cobbler is growing rather fast... VERY fast, and as it grows, I want to make sure the quality of the application everywhere is the best it can be.
There are a lot of things left undone, and various rough corners that still need to be polished. SUSE, Network Config, and Debian/Ubuntu support are all very recent and in varying degrees of being finished. We're not even packaged for all of them. There are still things that can't be done in the web application, etc. Lots of expansion, but we can do more to refine those things. (Does koan work on all of these platforms yet for virt AND replacement? Etc). If we expand too fast, we forget to go back and fill in that detail.
I seem to encounter a lot of users that /do/ clone physical windows systems, so I am not sure doing scripted installs is the answer. Further, the RIS stuff does not seem to be consistent across all OSes . An approach based on cloning those systems may be better in the end because that also allows working systems to be deployed .. not just empty shells that just have the Base OS on them.
I think we need to take a step back for a bit and focus on the core of Cobbler, close out lots of bugs and RFEs, ensure our new distribution support is first-class on the level with our others, and make sure the web application has parity with the command line.
Yes, all of these things /can/ be done in parallel, but as we add new arms to our octopus, we must make sure we go back and add the proverbial suckers to those arms.
Perfectionism is /not/ a goal, but I do want to make sure we don't spread out too fast and leave things unfinished.
Regarding the cloning of physical systems, a while back Andrew Brown contributed a live image that would clone (a bit slowly) a physical system using cobbler to network deploy the clone image. I think it might be interesting to try to resurrect and supercharge that idea. This way we could be sure to deploy Windows systems that /also/ had their software installed, without having to also teach Cobbler about Windows domain automation. We say Clonezilla and other analogs is hard to deploy, why not write a better solution?
And, of course, where virtualization is possible, I think that /IS/ a better route. We should encourage more use of that technology and also enhance our capabilities there where it makes sense.
--Michael
Michael-
What you layout is the best aproach, I work in many windows enviroments and RIS sucks, so we have always depend on imgaging to roll out new Windows desktops to servers. If the brialiant coders could figure out a way to have cobbler clone an existing system and redeploy it anyware, that would satisfy the Windows admins I know.
On Wed, Mar 25, 2009 at 8:17 AM, Rodrique Heron swygue@rodhouse.org wrote:
What you layout is the best aproach, I work in many windows enviroments and RIS sucks, so we have always depend on imgaging to roll out new Windows desktops to servers. If the brialiant coders could figure out a way to have cobbler clone an existing system and redeploy it anyware, that would satisfy the Windows admins I know.
Here is an insane idea for you to chew on. Perhaps someone could take the virt-p2v[1] livecd and make it smart enough to spit out a "cloneable" image that could be pxe-deployed by cobbler?
[1] http://et.redhat.com/~rjones/virt-p2v/
Jeff Schroeder wrote:
On Wed, Mar 25, 2009 at 8:17 AM, Rodrique Heron swygue@rodhouse.org wrote:
What you layout is the best aproach, I work in many windows enviroments and RIS sucks, so we have always depend on imgaging to roll out new Windows desktops to servers. If the brialiant coders could figure out a way to have cobbler clone an existing system and redeploy it anyware, that would satisfy the Windows admins I know.
Here is an insane idea for you to chew on. Perhaps someone could take the virt-p2v[1] livecd and make it smart enough to spit out a "cloneable" image that could be pxe-deployed by cobbler?
Need to look more at the internals of that.
What Andrew had was kind of like p2v ... used a live media to clone the disk image using partimage-ng to NFS, and the same image had a loader mode.
Check this out here... https://fedorahosted.org/cobbler/wiki/PhysicalSystemCloneFeature
How about we try to get this going again?
I think it mostly worked in his lab install... though I was seeing some issues with it... and we might want to try to speed it up or do something else slightly different using the same general idea.
There are a lot of smart folks on this list, I'm sure we can figure something out.
--Michael
On Wed, Mar 25, 2009 at 14:49, Michael DeHaan mdehaan@redhat.com wrote:
Robert Lazzurs wrote:
On Mon, Mar 23, 2009 at 16:40, Rainer Duffner rainer@ultra-secure.de wrote:
Michael DeHaan schrieb:
We don't really have a vested interest in making Windows physical deployments easy to deploy.
Hello,
So just to add my 2c in...
This statement worries me, I would suggest while we as a community should encourage the use of FOSS operating systems do we really want to take an RMS style view on this? Cobbler from my POV should be a generic systems management and will work best if it can document and deploy your whole infrastructure, re my earlier comments about upgrading iLO cards and network devices from Cobbler.
Windows is not going to be going away from the data centre any time soon and while I believe everything should be virtual unless there is a good reason not to be not everyone has that view.
Ok, I'll explain this further.
Cobbler is growing rather fast... VERY fast, and as it grows, I want to make sure the quality of the application everywhere is the best it can be.
There are a lot of things left undone, and various rough corners that still need to be polished. SUSE, Network Config, and Debian/Ubuntu support are all very recent and in varying degrees of being finished. We're not even packaged for all of them. There are still things that can't be done in the web application, etc. Lots of expansion, but we can do more to refine those things. (Does koan work on all of these platforms yet for virt AND replacement? Etc). If we expand too fast, we forget to go back and fill in that detail.
I seem to encounter a lot of users that /do/ clone physical windows systems, so I am not sure doing scripted installs is the answer. Further, the RIS stuff does not seem to be consistent across all OSes . An approach based on cloning those systems may be better in the end because that also allows working systems to be deployed .. not just empty shells that just have the Base OS on them.
I think we need to take a step back for a bit and focus on the core of Cobbler, close out lots of bugs and RFEs, ensure our new distribution support is first-class on the level with our others, and make sure the web application has parity with the command line.
Yes, all of these things /can/ be done in parallel, but as we add new arms to our octopus, we must make sure we go back and add the proverbial suckers to those arms.
Hello,
Agreed this is happening and does need to slow down a bit. Some crazy release manager around here keeps deciding we need to get those new features out yesterday ;)
I will do what I can to close more bugs and testing, however I have noticed that people are not closing bugs as they solve them. Please have a look over the bugs everyone, I am sure not every bug there should be assigned to mdehaan ;)
Now as for the platform testing, I have added a structure for the wiki for this, what do people think? Is this a worthwhile thing doing? https://fedorahosted.org/cobbler/wiki/PlatformTesting
So finally, once we have more of the "suckers on the arms" as you put it I presume as long as it does not become a support burden you have no objection to Cobbler being able to install Windows on bare tin...as long as the deployment method makes sense?
Take care.
Regarding the cloning of physical systems, a while back Andrew Brown contributed a live image that would clone (a bit slowly) a physical system using cobbler to network deploy the clone image. I think it might be interesting to try to resurrect and supercharge that idea. This way we could be sure to deploy Windows systems that /also/ had their software installed, without having to also teach Cobbler about Windows domain automation. We say Clonezilla and other analogs is hard to deploy, why not write a better solution?
We're in the process of getting clonezilla up & running. We've added it to cobbler as a distro, without any real issues, and are actually intending to use to for RHEL based systems, vs windows. We have a few Oracle RAC servers, and they're all identical to each other except for the host name, and a few settings that the DBA's need to adjust. It'll be much faster than using a kickstart to build the system, in our case.
We also made the Windows team in the office (vs the datacenter) happy, by allowing them to use the pxe to pull up clonezilla. They can reimage a machine without even moving it.
I'm not sure of what the issue was with clonezilla in the past, but it seems pretty straight forward to us.
Having said that - it *really* doesn't like booting with the fiber cards we have on our hardware, so.. i'll be looking into the live image for cloning. Maybe that works :-)
Matthew
------------------ Matthew Barr InteractiveOne - Senior System Engineer e:mbarr@interactiveone.com
Not that I understand the pitfalls of clonezilla, but this sounds like a good start. Has a RFE ticket been submitted yet for some sort of cloning? I didn't see one.
-----Original Message----- From: cobbler-bounces@lists.fedorahosted.org [mailto:cobbler- bounces@lists.fedorahosted.org] On Behalf Of Matthew Barr Sent: Friday, April 03, 2009 1:52 AM To: cobbler mailing list Subject: Re: Stategery regarding Windows support for mixed
environments
Regarding the cloning of physical systems, a while back Andrew Brown contributed a live image that would clone (a bit slowly) a physical system using cobbler to network deploy the clone image. I think it might be interesting to try to resurrect and supercharge that idea. This way we could be sure to deploy Windows systems that /also/ had their software installed, without having to also
teach
Cobbler about Windows domain automation. We say Clonezilla and other analogs is hard to deploy, why not write a better solution?
We're in the process of getting clonezilla up & running. We've added it to cobbler as a distro, without any real issues, and are actually intending to use to for RHEL based systems, vs windows. We have a few Oracle RAC servers, and they're all identical to each other except for the host name, and a few settings that the DBA's need to adjust. It'll be much faster than using a kickstart to build the system, in our case.
We also made the Windows team in the office (vs the datacenter) happy, by allowing them to use the pxe to pull up clonezilla. They can reimage a machine without even moving it.
I'm not sure of what the issue was with clonezilla in the past, but it seems pretty straight forward to us.
Having said that - it *really* doesn't like booting with the fiber cards we have on our hardware, so.. i'll be looking into the live image for cloning. Maybe that works :-)
Matthew
Matthew Barr InteractiveOne - Senior System Engineer e:mbarr@interactiveone.com
cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
On Mon, 23 Mar 2009 12:06:39 -0400, Michael DeHaan mdehaan@redhat.com wrote:
This is a development branch thing but it's related to Cobbler's plan moving forward, so it seems best to discuss here.
So ... there's been some good work going so far towards supporting some variants of Windows, though I see it may perhaps be somewhat of an evolutionary dead-end.
The web page for RIS-Linux in particular does not mention Vista or 2008. Further, I am much more interested in supporting Windows
virtually
than physically -- this should be natural with the increasing interest in virtualization and recent interoperability agreements with Microsoft.
For the devel branch (1.7), I'd rather we refocus our efforts into making sure the experience for Windows installs, virtually, is as good as possible.
The upside of this is that most of the infrastructure is already in place -- we can already do ISO based fullvirt installs ("cobbler image add" with the ISO file residing on NFS) and a next step is to teach it about a virtual floppy drive with the SIF answer file on the drive, so it can be fully scripted.
We already also have the "virt-clone" image type, for being able to take an existing disk image and repeatedly clone that image with koan, keeping the same source image on NFS. (The syntax here is "koan --image=foo --virt", just like with the ISO based installs for virt).
The goal here is to not invest too much effort in supporting dead-end deployment areas and writing code to cater to say, XP vs Vista vs 2003 vs 2008, but handle things generically, with answer files, and images, things we already do. On the plus side, there's also much less work in doing this and no additional dependencies or things to configure and set up.
Rather than physical deployments this encourages deploying Windows on Linux hosts, which makes the Windows machines easier to manage since you can run tools like libvirt and Func on the hosts. See http://www.redhat.com/promo/svvp/
Windows belongs running on virt. In general, for non-Linux OS's, we should also concentrate on virt.
--Michael
cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
Here are my thoughts on why this may be the correct long term goal, but bad for the short term (and by short term I mean the next 5 years). Small-Mid size corporations are quite often heterogenous, with many different OS'es running in the server room/data center. This is also true with Large companies, though you run into issues of momentum and territoriality there, so getting new build environments in is not an easy thing to do.
When I was working at a pretty good sized hosting company/Tier-1 ISP, one of the complaints I fielded against cobbler was that it was Red Hat specific, which at the time it very much was. In a company that deployed RH/Solaris/Windows with fairly equal regularity on real hardware, there would be no chance of replacing the in house build systems with cobbler.
Virtualization certainly has its place, however currently there are many pitfalls preventing its complete adoption (real-time systems in particular). As I stated above, I expect this to change within 5 years or so, but for now this decision would take away the ability of admins to use cobbler as the Swiss-army knife of build servers when deploying to bare metal systems. I also believe that if you can establish a beach head in the data center for cobbler, you will open avenues for adoption by demonstrating how easy it is to deploy virtually (no pun intended) as many OSes as are commonly used on many different platforms.
I'm obviously a bit biased, being the developer interested in supporting as many non-linux os'es as possible, although I certainly understand the RH viewpoint and the stategery involved in the decision. I won't be heart broken if we don't decide to support bare metal for most (if not all) non-linux os'es, but I do think it's a missed opportunity.
James Cammarata wrote:
On Mon, 23 Mar 2009 12:06:39 -0400, Michael DeHaan mdehaan@redhat.com wrote:
This is a development branch thing but it's related to Cobbler's plan moving forward, so it seems best to discuss here.
So ... there's been some good work going so far towards supporting some variants of Windows, though I see it may perhaps be somewhat of an evolutionary dead-end.
The web page for RIS-Linux in particular does not mention Vista or 2008. Further, I am much more interested in supporting Windows
virtually
than physically -- this should be natural with the increasing interest in virtualization and recent interoperability agreements with Microsoft.
For the devel branch (1.7), I'd rather we refocus our efforts into making sure the experience for Windows installs, virtually, is as good as possible.
The upside of this is that most of the infrastructure is already in place -- we can already do ISO based fullvirt installs ("cobbler image add" with the ISO file residing on NFS) and a next step is to teach it about a virtual floppy drive with the SIF answer file on the drive, so it can be fully scripted.
We already also have the "virt-clone" image type, for being able to take an existing disk image and repeatedly clone that image with koan, keeping the same source image on NFS. (The syntax here is "koan --image=foo --virt", just like with the ISO based installs for virt).
The goal here is to not invest too much effort in supporting dead-end deployment areas and writing code to cater to say, XP vs Vista vs 2003 vs 2008, but handle things generically, with answer files, and images, things we already do. On the plus side, there's also much less work in doing this and no additional dependencies or things to configure and set up.
Rather than physical deployments this encourages deploying Windows on Linux hosts, which makes the Windows machines easier to manage since you can run tools like libvirt and Func on the hosts. See http://www.redhat.com/promo/svvp/
Windows belongs running on virt. In general, for non-Linux OS's, we should also concentrate on virt.
--Michael
cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
Here are my thoughts on why this may be the correct long term goal, but bad for the short term (and by short term I mean the next 5 years). Small-Mid size corporations are quite often heterogenous, with many different OS'es running in the server room/data center. This is also true with Large companies, though you run into issues of momentum and territoriality there, so getting new build environments in is not an easy thing to do.
When I was working at a pretty good sized hosting company/Tier-1 ISP, one of the complaints I fielded against cobbler was that it was Red Hat specific, which at the time it very much was. In a company that deployed RH/Solaris/Windows with fairly equal regularity on real hardware, there would be no chance of replacing the in house build systems with cobbler.
Roger that.
It's increasingly becoming /Linux/ specific, not Red Hat specific, and agnostic to anything running inside a fully virtualized VM (ISO based FV installs, virt clones, etc).
In many cases we also see that the Windows side of the house is managed from some other tool, regardless, so we don't need to be able to replace that. When we do, we have virt as our sword -- it allows supporting *anything*.
Virtualization certainly has its place, however currently there are many pitfalls preventing its complete adoption (real-time systems in particular). As I stated above, I expect this to change within 5 years or so, but for now this decision would take away the ability of admins to use cobbler as the Swiss-army knife of build servers when deploying to bare metal systems. I also believe that if you can establish a beach head in the data center for cobbler, you will open avenues for adoption by demonstrating how easy it is to deploy virtually (no pun intended) as many OSes as are commonly used on many different platforms.
I'd rather seek to do what we do the best that we can do it, rather than spreading ourselves too thin. For mixed Linux/other environments the answer should be virt.
I'm obviously a bit biased, being the developer interested in supporting as many non-linux os'es as possible, although I certainly understand the RH viewpoint and the stategery involved in the decision. I won't be heart broken if we don't decide to support bare metal for most (if not all) non-linux os'es, but I do think it's a missed opportunity.
*nod*
As much as I'd like to, we can't hit all the opportunities.
I want to see virtualized Windows (and other OS's) advance, the way we do this is to avoid putting Windows physical installations on life support, and asking folks to embrace the virt tooling.
--Michael
-----Original Message----- From: cobbler-bounces@lists.fedorahosted.org [mailto:cobbler- bounces@lists.fedorahosted.org] On Behalf Of Michael DeHaan Sent: Monday, March 23, 2009 11:42 AM To: jimi@sngx.net; cobbler mailing list Subject: Re: Stategery regarding Windows support for mixed
environments
James Cammarata wrote:
On Mon, 23 Mar 2009 12:06:39 -0400, Michael DeHaan
wrote:
This is a development branch thing but it's related to Cobbler's
plan
moving forward, so it seems best to discuss here.
So ... there's been some good work going so far towards supporting
some
variants of Windows, though I see it may perhaps be somewhat of an evolutionary dead-end.
The web page for RIS-Linux in particular does not mention Vista or 2008. Further, I am much more interested in supporting Windows
virtually
than physically -- this should be natural with the increasing
interest
in virtualization and recent interoperability agreements with
Microsoft.
I did originally wonder why there was a focus on RIS and not WDS. Personally, that works well for me, because Vista is being avoided like the plague in many circles including my own company. I understand that focusing on re-implementing the M$ installation method du jour, could be frustrating, but it seemed that the idea was to leverage the ris-linux package which was already doing much of the work. I wouldn't expect a WDS update for cobbler until a wds-linux project is in a well-working state.
For the devel branch (1.7), I'd rather we refocus our efforts into making sure the experience for Windows installs, virtually, is as
good
as possible.
The upside of this is that most of the infrastructure is already in place -- we can already do ISO based fullvirt installs ("cobbler
image
add" with the ISO file residing on NFS) and a next step is to teach it about a virtual
floppy
drive with the SIF answer file on the drive, so it can be fully
scripted.
We already also have the "virt-clone" image type, for being able to
take
an existing disk image and repeatedly clone that image with koan, keeping the same source image on NFS. (The syntax here is "koan --image=foo
--virt",
just like with the ISO based installs for virt).
The goal here is to not invest too much effort in supporting dead-
end
deployment areas and writing code to cater to say, XP vs Vista vs
2003
vs 2008, but handle things generically, with answer files, and images, things we
already
do. On the plus side, there's also much less work in doing this
and
no additional dependencies or things to configure and set up.
Rather than physical deployments this encourages deploying Windows
on
Linux hosts, which makes the Windows machines easier to manage
since
you
can run tools like libvirt and Func on the hosts. See http://www.redhat.com/promo/svvp/
Windows belongs running on virt. In general, for non-Linux OS's,
we
should also concentrate on virt.
--Michael
cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
Here are my thoughts on why this may be the correct long term goal,
but bad
for the short term (and by short term I mean the next 5 years).
Small-Mid
size corporations are quite often heterogenous, with many different
OS'es
running in the server room/data center. This is also true with
Large
companies, though you run into issues of momentum and territoriality
there,
so getting new build environments in is not an easy thing to do.
When I was working at a pretty good sized hosting company/Tier-1
ISP,
one
of the complaints I fielded against cobbler was that it was Red Hat specific, which at the time it very much was. In a company that
deployed
RH/Solaris/Windows with fairly equal regularity on real hardware,
there
would be no chance of replacing the in house build systems with
cobbler.
Roger that.
It's increasingly becoming /Linux/ specific, not Red Hat specific, and agnostic to anything running inside a fully virtualized VM (ISO based FV installs, virt clones, etc).
In many cases we also see that the Windows side of the house is
managed
from some other tool, regardless, so we don't need to be able to replace that. When we do, we have virt as our sword -- it allows supporting *anything*.
Virtualization certainly has its place, however currently there are
many
pitfalls preventing its complete adoption (real-time systems in particular). As I stated above, I expect this to change within 5
years or
so, but for now this decision would take away the ability of admins
to use
cobbler as the Swiss-army knife of build servers when deploying to
bare
metal systems. I also believe that if you can establish a beach
head
in
the data center for cobbler, you will open avenues for adoption by demonstrating how easy it is to deploy virtually (no pun intended)
as
many
OSes as are commonly used on many different platforms.
I agree, the Linux philosophy definitely favors the Swiss-army knife approach. It may not be the easiest OS for joe-schmoe to pick up and start using, it may not be the cleanest user experience for the novice user, but it can do whatever you want it to do.
I'd rather seek to do what we do the best that we can do it, rather than spreading ourselves too thin. For mixed Linux/other environments the answer should be virt.
OK, I know you didn't mean it this way, but you telling me what the answer is before I've asked the question is starting to sound more like M$ itself. I'm sure for your deployment Windows run virtually on top of Linux works great. For my deployment, it's a complete non-starter. Vendors that we buy software from are not going to support their software if it's running in this manner. This is especially true for us since most of our software must interoperate with hardware (via USB, firewire, Ethernet, RS232, and proprietary PCI cards to name a few).
I'm obviously a bit biased, being the developer interested in
supporting as
many non-linux os'es as possible, although I certainly understand
the
RH
viewpoint and the stategery involved in the decision. I won't be
heart
broken if we don't decide to support bare metal for most (if not
all)
non-linux os'es, but I do think it's a missed opportunity.
*nod*
As much as I'd like to, we can't hit all the opportunities.
I want to see virtualized Windows (and other OS's) advance, the way we do this is to avoid putting Windows physical installations on life support, and asking folks to embrace
the
virt tooling.
As I mentioned, my company would embrace this about as well as it's embraced Vista (i.e. not at all).
--Michael
cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
If you're totally against the RIS integration, what about beefing up the documentation and interfacing for a secondary installation server to install the non-Linux systems? Seems like there could be non-local profiles that point to another system via tftp and grub.
As I mentioned, my company would embrace this about as well as it's embraced Vista (i.e. not at all).
Hi Jared,
We want to look into integration with cloning tools, or resurrection of the cloning live image that we had going several months ago for physical installs. Being able to have all the software installed "full stack" is important, and the workflow will be smoother to get all the software onto the box -- while also straining development time much less and offering solutions for all OSes, not just Windows. (Yes, you could use this to image GNU Hurd if you wanted).
Meanwhile, start that Linux port :)
--Michael
-----Original Message----- From: cobbler-bounces@lists.fedorahosted.org [mailto:cobbler- bounces@lists.fedorahosted.org] On Behalf Of Michael DeHaan Sent: Wednesday, April 01, 2009 7:59 AM To: cobbler mailing list Subject: Re: Stategery regarding Windows support for mixed
environments
As I mentioned, my company would embrace this about as well as it's embraced Vista (i.e. not at all).
Hi Jared,
We want to look into integration with cloning tools, or resurrection
of
the cloning live image that we had going several months ago for physical installs. Being able to have all the software installed "full
stack"
is important, and the workflow will be smoother to get all the
software
onto the box -- while also straining development time much less and offering solutions for all OSes, not just Windows. (Yes, you could use this to image GNU Hurd if you wanted).
Meanwhile, start that Linux port :)
--Michael _______________________________________________ cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
Michael, The answer of "cloning" seems vastly superior to the "you should use virtualization". I agree "full stack" is important. We currently use some form of ghost for Windows systems (as we're still trying to get cobbler up and running for WinXP installs via ris-linux). I'd encourage the use of some sort of intelligent "ghost" package, and not something very low-level (dd anyone?), so reinventing the advanced functionality is not necessary. Thanks.
-Jared
Hi,
On Mon, Mar 23, 2009 at 12:06:39PM -0400, Michael DeHaan wrote:
Windows belongs running on virt. In general, for non-Linux OS's, we should also concentrate on virt.
there is still *new* hardware out there not supporting hardware virtualization which Windows seems to require.
Currently I'm looking at a small sized deployment at a school where no system supports hardware virtualization and I assume that most other schools don't either.
Just a data point.
Axel Thimm wrote:
Hi,
On Mon, Mar 23, 2009 at 12:06:39PM -0400, Michael DeHaan wrote:
Windows belongs running on virt. In general, for non-Linux OS's, we should also concentrate on virt.
there is still *new* hardware out there not supporting hardware virtualization which Windows seems to require.
Currently I'm looking at a small sized deployment at a school where no system supports hardware virtualization and I assume that most other schools don't either.
Just a data point.
Absolutely, you have to plan for it when you are building out your environment. More so for schools, they are likely to have old, if not ancient, hardware.
Possibly http://clonezilla.org/ ?
--Michael
On Tue, Mar 24, 2009 at 01:50:39PM -0400, Michael DeHaan wrote:
Axel Thimm wrote:
Hi,
On Mon, Mar 23, 2009 at 12:06:39PM -0400, Michael DeHaan wrote:
Windows belongs running on virt. In general, for non-Linux OS's, we should also concentrate on virt.
there is still *new* hardware out there not supporting hardware virtualization which Windows seems to require.
Currently I'm looking at a small sized deployment at a school where no system supports hardware virtualization and I assume that most other schools don't either.
Just a data point.
Absolutely, you have to plan for it when you are building out your environment. More so for schools, they are likely to have old, if not ancient, hardware.
Possibly http://clonezilla.org/ ?
I've looked at it, especially the parts that try to restore hostname/sid information. It does look a bit cumbersome to setup and I manage similar things with booting linux off pxe and using ntfsclone.
Is it possible to boot into cobbler/anaconda and skip the package selection part? E.g. only have custom scripts %post applied like ntfsclone from an nfs share?
On Tue, Mar 24, 2009 at 11:53 AM, Axel Thimm Axel.Thimm@atrpms.net wrote:
On Tue, Mar 24, 2009 at 01:50:39PM -0400, Michael DeHaan wrote:
Axel Thimm wrote:
Hi,
On Mon, Mar 23, 2009 at 12:06:39PM -0400, Michael DeHaan wrote:
Windows belongs running on virt. In general, for non-Linux OS's, we should also concentrate on virt.
there is still *new* hardware out there not supporting hardware virtualization which Windows seems to require.
Currently I'm looking at a small sized deployment at a school where no system supports hardware virtualization and I assume that most other schools don't either.
Just a data point.
Absolutely, you have to plan for it when you are building out your environment. More so for schools, they are likely to have old, if not ancient, hardware.
Possibly http://clonezilla.org/ ?
I've looked at it, especially the parts that try to restore hostname/sid information. It does look a bit cumbersome to setup and I manage similar things with booting linux off pxe and using ntfsclone.
Is it possible to boot into cobbler/anaconda and skip the package selection part? E.g. only have custom scripts %post applied like ntfsclone from an nfs share?
You might be able to hack this with something like this: %packages --nodocs --nobase
%post ...
Not sure if anaconda will barf but that might do what you want. Can you try it and let us know?
On Tue, Mar 24, 2009 at 12:02:41PM -0700, Jeff Schroeder wrote:
Is it possible to boot into cobbler/anaconda and skip the package selection part? E.g. only have custom scripts %post applied like ntfsclone from an nfs share?
You might be able to hack this with something like this: %packages --nodocs --nobase
%post ...
Not sure if anaconda will barf but that might do what you want. Can you try it and let us know?
Thanks, I'll try that! I can't promise a fast feedback on this as daywork has currently taken over nights as well.
Axel Thimm wrote:
On Tue, Mar 24, 2009 at 01:50:39PM -0400, Michael DeHaan wrote:
Axel Thimm wrote:
Hi,
On Mon, Mar 23, 2009 at 12:06:39PM -0400, Michael DeHaan wrote:
Windows belongs running on virt. In general, for non-Linux OS's, we should also concentrate on virt.
there is still *new* hardware out there not supporting hardware virtualization which Windows seems to require.
Currently I'm looking at a small sized deployment at a school where no system supports hardware virtualization and I assume that most other schools don't either.
Just a data point.
Absolutely, you have to plan for it when you are building out your environment. More so for schools, they are likely to have old, if not ancient, hardware.
Possibly http://clonezilla.org/ ?
I've looked at it, especially the parts that try to restore hostname/sid information. It does look a bit cumbersome to setup and I manage similar things with booting linux off pxe and using ntfsclone.
Yes, my thoughts were the same WRT setup difficulty.
Is it possible to boot into cobbler/anaconda and skip the package selection part? E.g. only have custom scripts %post applied like ntfsclone from an nfs share?
One way to do this is to use cobbler to PXE boot a live image. The live image can read any data it needs from /proc/cmdline (which can be passed via cobbler).
cobbler mailing list cobbler@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/cobbler
cobbler@lists.fedorahosted.org
-
Axel Thimm -
James Cammarata -
Jared Hodge -
Jeff Schroeder -
Matthew Barr -
Michael DeHaan -
Rainer Duffner -
Robert Lazzurs -
Rodrique Heron