hi guys,
lots of people on this list are keen on contributing to the CentOS effort - and one place where we could, as a group of people, make a massive difference is being able to deliver a ksplice upgrade model. A couple of people have started in silo's to try and get something together but I feel it might be a good time to try and see if we can do somethng as a group.
Do we have people on the list who find the idea of rebootless kernel upgrades interesting ? Is there some desire to try and make that happen within the CentOS Ecosystem ?
- KB
On Thu, Jul 21, 2011 at 3:37 PM, Karanbir Singh mail-lists@karan.org wrote:
hi guys,
lots of people on this list are keen on contributing to the CentOS effort - and one place where we could, as a group of people, make a massive difference is being able to deliver a ksplice upgrade model. A couple of people have started in silo's to try and get something together but I feel it might be a good time to try and see if we can do somethng as a group.
Do we have people on the list who find the idea of rebootless kernel upgrades interesting ? Is there some desire to try and make that happen within the CentOS Ecosystem ?
So you mean, developing from the ground up, something "ksplice-like"? The "ksplice" I see out there is closed source, and was purchased by Oracle today, for what it's worth... :-(
jerry
On Thu, Jul 21, 2011 at 03:56:48PM -0500, Jerry Amundson wrote:
So you mean, developing from the ground up, something "ksplice-like"? The "ksplice" I see out there is closed source, and was purchased by Oracle today, for what it's worth... :-(
And in typical Oracraple tradition immediately dropped support for RHEL and Suse.
John
On Thu, Jul 21, 2011 at 04:03:41PM -0500, John R. Dennison wrote:
On Thu, Jul 21, 2011 at 03:56:48PM -0500, Jerry Amundson wrote:
So you mean, developing from the ground up, something "ksplice-like"? The "ksplice" I see out there is closed source, and was purchased by Oracle today, for what it's worth... :-(
And in typical Oracraple tradition immediately dropped support for RHEL and Suse.
Sorry to reply to my own message but they've also removed CentOS from the list of supported platforms.
John
On 21 July 2011 22:09, John R. Dennison jrd@gerdesas.com wrote:
Sorry to reply to my own message but they've also removed CentOS from the list of supported platforms.
Do I live in a parallel universe? http://www.ksplice.com/uptrack/supported-kernels has
Ksplice Uptrack supported kernels Ksplice Uptrack can bring the following kernels up to date with the latest important security and bug fix patches:
Red Hat and CentOS All CentOS and RHEL 6 kernels starting with with the official release All CentOS and RHEL 5 non-Xen kernels starting with 2.6.18-92.1.22.el5 (released Dec. 5, 2008) All CentOS and RHEL 5 Xen kernels starting with 2.6.18-128.1.14.el5xen (released June 1, 2009) All CentOSPlus 5 non-Xen kernels starting with 2.6.18-92.1.22.el5.centos.plus (released Dec. 17, 2008) All CentOSPlus 5 Xen kernels starting with 2.6.18-128.1.14.el5.centos.plusxen (released June 20, 2009) All CentOS and RHEL 4 non-Xen kernels starting with 2.6.9-67.EL (released Nov. 15, 2007) All CentOSPlus 4 "plus.c4" and "plus.c4smp" kernels starting with 2.6.9-78.0.8.EL (released Nov. 5, 2008)
and so on?
On Thu, Jul 21, 2011 at 10:18:44PM +0100, Hakan Koseoglu wrote:
On 21 July 2011 22:09, John R. Dennison jrd@gerdesas.com wrote:
Sorry to reply to my own message but they've also removed CentOS from the list of supported platforms.
Do I live in a parallel universe? http://www.ksplice.com/uptrack/supported-kernels has
Ksplice Uptrack supported kernels Ksplice Uptrack can bring the following kernels up to date with the latest important security and bug fix patches:
but the header reads: On July 21, 2011, Oracle announced that it has acquired Ksplice. ...` points to http://www.oracle.com/us/corporate/Acquisitions/ksplice/index.html and http://www.oracle.com/us/corporate/press/435791 ... Oracle believes it will be the only enterprise Linux provider that can offer zero downtime updates, and expects to make the Ksplice technology a standard feature of Oracle Linux Premier Support. ...
Time will tell
Tru
On 21/07/11 22:29, Tru Huynh wrote:
Oracle believes it will be the only enterprise Linux provider that can offer zero downtime updates, and expects to make the Ksplice technology a standard feature of Oracle Linux Premier Support. ...
Time will tell
Ho-hum. We'll see. Oracle is becoming one of the most evil companies as far as my opinion counts.
On Fri, Jul 22, 2011 at 3:22 AM, Karanbir Singh mail-lists@karan.orgwrote:
On 07/21/2011 10:39 PM, Hakan Koseoglu wrote:
Ho-hum. We'll see. Oracle is becoming one of the most evil companies as far as my opinion counts.
Sounds like a good reason for us to pickup on ksplice and see if we, as a group of people, can create an alternative.
- KB
+1 Please proceed ahead. This is one of the most interesting stuff - applying most of the security patches without rebooting on production servers. Oracle has not followed the community tradition and collectively we can bring back this exiting stuff to mainstream Linux.
Karanbir Singh wrote:
So, what are you going to be doing to help?
It would be good to learn how exactly would development go.
- What is programing language, - what are current caveats, bugs - what are the goals for the future, - and what is estimated number of developers and testers required/desired.
That would help people decide what they could offer.
On Thu, Jul 21, 2011 at 11:18 PM, Hakan Koseoglu hakan@koseoglu.org wrote:
Do I live in a parallel universe? http://www.ksplice.com/uptrack/supported-kernels has
Ksplice Uptrack supported kernels Ksplice Uptrack can bring the following kernels up to date with the latest important security and bug fix patches:
Red Hat and CentOS All CentOS and RHEL 6 kernels starting with with the official release All CentOS and RHEL 5 non-Xen kernels starting with 2.6.18-92.1.22.el5 (released Dec. 5, 2008) All CentOS and RHEL 5 Xen kernels starting with 2.6.18-128.1.14.el5xen (released June 1, 2009) All CentOSPlus 5 non-Xen kernels starting with 2.6.18-92.1.22.el5.centos.plus (released Dec. 17, 2008) All CentOSPlus 5 Xen kernels starting with 2.6.18-128.1.14.el5.centos.plusxen (released June 20, 2009) All CentOS and RHEL 4 non-Xen kernels starting with 2.6.9-67.EL (released Nov. 15, 2007) All CentOSPlus 4 "plus.c4" and "plus.c4smp" kernels starting with 2.6.9-78.0.8.EL (released Nov. 5, 2008)
and so on?
yes but then in download links you can find only Fedora and Ubuntu. Also if you select "trial ksplice" only these two distributions ... while before you could try RH EL and CentOS too and evaluate the product before buy. In fact I tried it when there was a serious bug in kernel with root privileges escalation (CVE-2010-3081) and an open exploit and verified its goodness. So +1 for me
Gianluca
On 07/21/2011 09:56 PM, Jerry Amundson wrote:
So you mean, developing from the ground up, something "ksplice-like"? The "ksplice" I see out there is closed source, and was purchased by Oracle today, for what it's worth... :-(
ksplice has always been GPL friendly, or I've been lead to believe by the guys who wrote it originally. I have a git clone from earlier this year and it clearly indicates the code as being GPLv2.
Happy to publish it.
- KB
On Thu, Jul 21, 2011 at 4:04 PM, Karanbir Singh mail-lists@karan.org wrote:
On 07/21/2011 09:56 PM, Jerry Amundson wrote:
So you mean, developing from the ground up, something "ksplice-like"? The "ksplice" I see out there is closed source, and was purchased by Oracle today, for what it's worth... :-(
ksplice has always been GPL friendly, or I've been lead to believe by the guys who wrote it originally. I have a git clone from earlier this year and it clearly indicates the code as being GPLv2.
Happy to publish it.
Please do! :-) The download links in this WikiPedia article are already zapped as well. http://en.wikipedia.org/wiki/Ksplice
jerry
On Thu, Jul 21, 2011 at 04:10:10PM -0500, Jerry Amundson wrote:
On Thu, Jul 21, 2011 at 4:04 PM, Karanbir Singh mail-lists@karan.org wrote:
On 07/21/2011 09:56 PM, Jerry Amundson wrote:
So you mean, developing from the ground up, something "ksplice-like"? The "ksplice" I see out there is closed source, and was purchased by Oracle today, for what it's worth... :-(
ksplice has always been GPL friendly, or I've been lead to believe by the guys who wrote it originally. I have a git clone from earlier this year and it clearly indicates the code as being GPLv2.
Happy to publish it.
Please do! :-) The download links in this WikiPedia article are already zapped as well. http://en.wikipedia.org/wiki/Ksplice
http://dev.centos.org/~tru/ksplice/ original files and signed with their key
from the included README:
Ksplice Copyright (C) 2007-2009 Ksplice, Inc. Authors: Jeff Arnold, Anders Kaseorg, Tim Abbott
This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License, version 2.
For more information about Ksplice, please see http://www.ksplice.com/.
Installation instructions: $ ./configure $ make $ make install
Notable Build Dependencies: - GNU binary file descriptor (BFD) library (version 2.15 or later) (available in Debian's binutils-dev package and in other distributions)
Tru
On Friday 22 July 2011 00:25:29 Tru Huynh wrote:
On Thu, Jul 21, 2011 at 04:10:10PM -0500, Jerry Amundson wrote:
On Thu, Jul 21, 2011 at 4:04 PM, Karanbir Singh mail-lists@karan.org
wrote:
On 07/21/2011 09:56 PM, Jerry Amundson wrote:
So you mean, developing from the ground up, something "ksplice-like"? The "ksplice" I see out there is closed source, and was purchased by Oracle today, for what it's worth... :-(
ksplice has always been GPL friendly, or I've been lead to believe by the guys who wrote it originally. I have a git clone from earlier this year and it clearly indicates the code as being GPLv2.
Happy to publish it.
Please do! :-) The download links in this WikiPedia article are already zapped as well. http://en.wikipedia.org/wiki/Ksplice
http://dev.centos.org/~tru/ksplice/ original files and signed with their key
from the included README:
Ksplice Copyright (C) 2007-2009 Ksplice, Inc. Authors: Jeff Arnold, Anders Kaseorg, Tim Abbott
This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License, version 2.
For more information about Ksplice, please see http://www.ksplice.com/.
Installation instructions: $ ./configure $ make $ make install
Notable Build Dependencies:
- GNU binary file descriptor (BFD) library (version 2.15 or later) (available in Debian's binutils-dev package and in other distributions)
Tru
I have worked for a few days on the KSplice code. But I'm stuck writing a tool for analyzing patches.
The problem I faced was that if you have a few simple patches that don't alter major structures in the kernel it self you can simply create a ksplice module from those patches and load it. It works like a charm.
But when you start putting more and more patches, inevidablly you reach a point where there is a patch that for example changes the VFS layer or the network layer or somthing like that. At that point you need to write additional code so the ksplice module generator would know how to handle the upgrade.
For some patches it is fairly trivial... for others it is quite complex.
So what I decided to do is firs separate the 3 types of patches: 1. Ready to apply (no major changes to structures in the kernel) 2. Patches that require additional code (minor or trivial changes to structures in the kernel) 3. Patches that require manual review (major changes to structures in the kernel or possible conflicts with other parts of the kernel)
It would be great if we put the project in GitHub or Gitorius so we can all collaborate on the solution.
Best regards, Marian Marinov CEO of 1H Ltd.
Karanbir Singh wrote:
On 07/21/2011 11:09 PM, Marian Marinov wrote:
It would be great if we put the project in GitHub or Gitorius so we can all collaborate on the solution.
I've literally just put the git repo together and am pushing it out with a few mirrors. Stay tuned for urls
- KB
CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
Will you have to change a name for a fork? Maybe "cksplice"?
Ok, now with some git, you can get the base code from :
git clone git://nazar.karan.org/testing/ksplice
web ui at : https://nazar.karan.org/cgit/ksplice/
with mirror's at
https://gitorious.org/centos/ksplice and https://github.com/CentOS/ksplice
merge across mirrors via hooks ( lets see how that works out.. )
push changes via github / gitorious or git-email..
- KB
On 7/23/2011 3:29 PM, Karanbir Singh wrote:
Ok, now with some git, you can get the base code from :
git clone git://nazar.karan.org/testing/ksplice
web ui at : https://nazar.karan.org/cgit/ksplice/
with mirror's at
https://gitorious.org/centos/ksplice and https://github.com/CentOS/ksplice
merge across mirrors via hooks ( lets see how that works out.. )
push changes via github / gitorious or git-email..
- KB
CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
kewlies..glad to see somebody had the code too..:)
On 7/21/2011 4:56 PM, Jerry Amundson wrote:
On Thu, Jul 21, 2011 at 3:37 PM, Karanbir Singhmail-lists@karan.org wrote:
hi guys,
lots of people on this list are keen on contributing to the CentOS effort - and one place where we could, as a group of people, make a massive difference is being able to deliver a ksplice upgrade model. A couple of people have started in silo's to try and get something together but I feel it might be a good time to try and see if we can do somethng as a group.
Do we have people on the list who find the idea of rebootless kernel upgrades interesting ? Is there some desire to try and make that happen within the CentOS Ecosystem ?
So you mean, developing from the ground up, something "ksplice-like"? The "ksplice" I see out there is closed source, and was purchased by Oracle today, for what it's worth... :-(
jerry _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
actually ksplice .9 was gpl2. I snagged a copy if anyone wants it i'll stick it up on my website.
On 21/07/11 21:37, Karanbir Singh wrote:
Do we have people on the list who find the idea of rebootless kernel upgrades interesting ? Is there some desire to try and make that happen within the CentOS Ecosystem ?
Yes I do find them interesting, thanks for this. I have never used it myself though as it seemed to be only Ubuntu that supported out of the box and I haven't used Ubuntu for a while now. It would be nice to be able to do rebootless upgrades in CentOS, yes. Not sure how I could contribute though.
I too am sad to see KSplice gobbled up by (evil corporation here). Though I am in the same boat as Khusro in that I probably can't really contribute to active development I am sure I have some hardware that can be dedicated to testing and QA for a KSplice fork.
Cheers,
Sam
----- Original Message ----- From: "Khusro Jaleel" mailing-lists@kerneljack.com To: centos-devel@centos.org Sent: Friday, July 22, 2011 8:49:52 AM Subject: Re: [CentOS-devel] ksplice and CentOS
On 21/07/11 21:37, Karanbir Singh wrote:
Do we have people on the list who find the idea of rebootless kernel upgrades interesting ? Is there some desire to try and make that happen within the CentOS Ecosystem ?
Yes I do find them interesting, thanks for this. I have never used it myself though as it seemed to be only Ubuntu that supported out of the box and I haven't used Ubuntu for a while now. It would be nice to be able to do rebootless upgrades in CentOS, yes. Not sure how I could contribute though. _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
On Thu, Jul 21, 2011 at 4:27 PM, Sam Wilson kahn@the-mesh.org wrote:
I too am sad to see KSplice gobbled up by (evil corporation here). Though I am in the same boat as Khusro in that I probably can't really contribute to active development I am sure I have some hardware that can be dedicated to testing and QA for a KSplice fork.
Cheers,
Sam
I would not be surprised if we see RH do something about this. They contribute the most to linux development of anybody. Anybody know someone at RH to see if there are plans? Possibly done on the Fedora side and trickle down to RHEL/CentOS?
Brandon
On Jul 21, 2011, at 7:29 PM, Brandon Ooi wrote:
On Thu, Jul 21, 2011 at 4:27 PM, Sam Wilson kahn@the-mesh.org wrote: I too am sad to see KSplice gobbled up by (evil corporation here). Though I am in the same boat as Khusro in that I probably can't really contribute to active development I am sure I have some hardware that can be dedicated to testing and QA for a KSplice fork.
Cheers,
Sam
I would not be surprised if we see RH do something about this. They contribute the most to linux development of anybody. Anybody know someone at RH to see if there are plans? Possibly done on the Fedora side and trickle down to RHEL/CentOS?
Um, does it matter what Oracle and RedHat are doing?
This is Centos, most users are here because they cannot afford to pay for either RedHat or Oracle product for whatever reason.
To paraphrase a recent LWN quote-of-the-week (and not from smooge ;-)
If you're not paying for THEIR product, you ARE their product.
If you would like be "product", while have fun!
Meanwhile if splice is useful and their is interest, you're gonna have to generate a kernel patch stream somehow in order to make splice Just Work (and continue to work) in CentOS.
73 de Jeff
On Thu, 21 Jul 2011, Ray Van Dolson wrote:
On Thu, Jul 21, 2011 at 07:37:16PM -0400, Jeff Johnson wrote:
Um, does it matter what Oracle and RedHat are doing?
Only to the extent that without Red Hat, CentOS doesn't exist! :-)
umm ... CentOS is of course an outgrowth sub-project from cAos [admittedly, the poster child success story], which pre-dated RHEL rebuilds, just fine, thank you very much. Founded on discussions of: What if there was no Red Hat Linux
Spinning stabilized enterprise grade distributions is largely a matter of a willingness to lag in a stable branch, and to backport security matters as needed. Boring stuff
The FOSS community predated the upstream's product that CentOS rebuilds sources of ... and the wells of those sources (although partially fed from efforts of the upstream) will continue if the upstream did not exist ...
In IRC today, someone wandered into the main channel looking for a 'later and greater' ImageMagic ... looking at the logs, it took me a bit under 40 minutes (1731 to 1809) to 'freshen' this leaf node package and all BR and R dependencies to a very recent product
How many leaf node package fruit does one really need and use daily, anyway? In the most recent 1000 lines of my history on the laptop I am composing this on, I find:
$ history | awk {'print $2'} | sort | uniq -c | sort -n | wc
is only 85 commands. and most of those are unremarkable or part of bash, or local scriptlets
2 display 2 for 2 kill 2 ks 2 lynx 2 /sbin/ifconfig 2 su 2 thunar 2 /usr/bin/firefox 2 wc 3 echo 3 grep 3 lynxdump 3 ./maple-ide 3 scp 3 unzip 4 eject 4 pwd 4 tail 5 chmod 5 convert 5 history 5 which 6 arduino 6 ./README 6 rpm 7 firefox-charlie 8 gen-pw.sh 8 man 9 ps 11 cp 15 cat 18 mv 18 rm 19 xpdf 22 rpmbuild 27 sudo 28 backup-home-to-net.sh 31 mkdir 38 wget 41 startx 46 less 70 joe 128 ssh 139 ls 190 cd 1000 3111 26659
[I see ImageMagic primitives in that list a couple of times; 'joe' and 'xpdf' are gone from C6, but it does not matter, as I've built them locally for a long time to meet my habituations]
It might not be comfortable for a few months if we needed to live like that transition of no RHL to something, but not the end of the world for people not addicted to the latest and greatest
-- Russ herrold
Jeff Johnson wrote:
On Jul 21, 2011, at 7:29 PM, Brandon Ooi wrote:
On Thu, Jul 21, 2011 at 4:27 PM, Sam Wilson <kahn@the-mesh.org mailto:kahn@the-mesh.org> wrote:
I too am sad to see KSplice gobbled up by (evil corporation here). Though I am in the same boat as Khusro in that I probably can't really contribute to active development I am sure I have some hardware that can be dedicated to testing and QA for a KSplice fork. Cheers, SamI would not be surprised if we see RH do something about this. They contribute the most to linux development of anybody. Anybody know someone at RH to see if there are plans? Possibly done on the Fedora side and trickle down to RHEL/CentOS?
Um, does it matter what Oracle and RedHat are doing?
This is Centos, most users are here because they cannot afford to pay for either RedHat or Oracle product for whatever reason.
To paraphrase a recent LWN quote-of-the-week (and not from smooge ;-)
If you're not paying for THEIR product, you ARE their product.
If you would like be "product", while have fun!
Meanwhile if splice is useful and their is interest, you're gonna have to generate a kernel patch stream somehow in order to make splice Just Work (and continue to work) in CentOS.
73 de Jeff
There are CentOSPlus kernels available. They can be modified as desired.
Hi,
On 07/22/2011 12:29 AM, Brandon Ooi wrote:
I would not be surprised if we see RH do something about this. They contribute the most to linux development of anybody. Anybody know someone at RH to see if there are plans? Possibly done on the Fedora side and trickle down to RHEL/CentOS?
As Jeff and Russ already pointed out - it would be interesting to see what Red Hat do at their end, or even if they consider this sort of a deployment scenario as something their customers would be keen on. However, we are quite a large community here within the CentOS arena, and given that there are quite a few people who use CentOS, who are also insterested in this sort of a scenario - now is a good opportunity to put together some efforts and see if we can solve this problem, for ourselves.
In terms of people who cant help : Maybe we can get Marian to host a webinar or something of that nature that people interested could attend and see for themselves what the implications are, how the work might be done and what is needed to make things happen. I would be happy to organise something of this nature, should Marian agree.
- KB
On Friday 22 July 2011 14:50:20 Karanbir Singh wrote:
Hi,
On 07/22/2011 12:29 AM, Brandon Ooi wrote:
I would not be surprised if we see RH do something about this. They contribute the most to linux development of anybody. Anybody know someone at RH to see if there are plans? Possibly done on the Fedora side and trickle down to RHEL/CentOS?
As Jeff and Russ already pointed out - it would be interesting to see what Red Hat do at their end, or even if they consider this sort of a deployment scenario as something their customers would be keen on. However, we are quite a large community here within the CentOS arena, and given that there are quite a few people who use CentOS, who are also insterested in this sort of a scenario - now is a good opportunity to put together some efforts and see if we can solve this problem, for ourselves.
In terms of people who cant help : Maybe we can get Marian to host a webinar or something of that nature that people interested could attend and see for themselves what the implications are, how the work might be done and what is needed to make things happen. I would be happy to organise something of this nature, should Marian agree.
- KB
The next three weeks I will be extremely busy around my company, so I would not be able before 14.Aug. But after that I'll be more then happy to host a webinar on ksplice and what I already have tested.
Best regards, Marian Marinov CEO of 1H Ltd.
Hi,
How can I help?
I can do QA/testing on my hardware, but all six servers I could use for tests are all on C5 aboutand I'm not quite ready to move them on to C6 so I could have no more than two C6 boxes now.
Are you going to do a single ksplice patch out of an entire upstream vendor's C6 patchset?
Best regards, Dmitry Mikhailov
Great, I would be happy to listen!
Regards, Marko
On Friday 22 July 2011 14:50:20 Karanbir Singh wrote:
Hi,
On 07/22/2011 12:29 AM, Brandon Ooi wrote:
I would not be surprised if we see RH do something about this. They contribute the most to linux development of anybody. Anybody know someone at RH to see if there are plans? Possibly done on the Fedora side and trickle down to RHEL/CentOS?
As Jeff and Russ already pointed out - it would be interesting to see what Red Hat do at their end, or even if they consider this sort of a deployment scenario as something their customers would be keen on. However, we are quite a large community here within the CentOS arena, and given that there are quite a few people who use CentOS, who are also insterested in this sort of a scenario - now is a good opportunity to put together some efforts and see if we can solve this problem, for ourselves.
In terms of people who cant help : Maybe we can get Marian to host a webinar or something of that nature that people interested could attend and see for themselves what the implications are, how the work might be done and what is needed to make things happen. I would be happy to organise something of this nature, should Marian agree.
- KB
The next three weeks I will be extremely busy around my company, so I would not be able before 14.Aug. But after that I'll be more then happy to host a webinar on ksplice and what I already have tested.
Best regards, Marian Marinov CEO of 1H Ltd. _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
On Thu, 21 Jul 2011 21:37:26 +0100 Karanbir Singh mail-lists@karan.org wrote:
hi guys,
lots of people on this list are keen on contributing to the CentOS effort - and one place where we could, as a group of people, make a massive difference is being able to deliver a ksplice upgrade model. A couple of people have started in silo's to try and get something together but I feel it might be a good time to try and see if we can do somethng as a group.
Do we have people on the list who find the idea of rebootless kernel upgrades interesting ? Is there some desire to try and make that happen within the CentOS Ecosystem ?
guys, and gals? :-)
They recently introduced something called kexec-tools in Debian (testing /wheezy) and for those in the linux mint debian arena (LMDE), it didn't go well. I'm currently using Debian testing (not mintified) and don't have it installed. Here's the description, in case it's helpful:
This tool is used to load a kernel in memory and reboot into the kernel loaded in memory using the kexec system call.
Here's a link: http://packages.debian.org/testing/admin/kexec-tools
Cia W
On Friday 22 July 2011 07:12, Cia Watson wrote:
This tool is used to load a kernel in memory and reboot into the kernel loaded in memory using the kexec system call.
FIrst of all, kexec is not ksplice. kexec allows to reboot kernel without rebooting hardware, but it's still a reboot and a usual disruption of services associated with it. ksplice is a quite another thing. It allows to apply patches to an already running kernel with no disruption of services.
Second, kexec is included in CentOS already: [tn@ibm2 cropgui]$ rpm -q kexec-tools kexec-tools-1.102pre-96.el5_5.4
Not sure whether default CentOS kernels support kexec system call.
Best regards, Dmitry Mikhailov
I spoke with my colleges about Ksplice and one very interesting question popped up:
What will happen with our fork efforts if Oracle suddenly patent the Ksplice software ?
I know that in Europe we don't have the problem with software pattents and the GPL will protect us from any issues like that.
However what is the situation in the US and other countries that have software pattents ?
Marian
Marian Marinov wrote:
I spoke with my colleges about Ksplice and one very interesting question popped up:
What will happen with our fork efforts if Oracle suddenly patent the Ksplice software ?
I know that in Europe we don't have the problem with software pattents and the GPL will protect us from any issues like that.
However what is the situation in the US and other countries that have software pattents ?
Marian
Any code released as GPL can be forked as long as fork is also GPL. Can Oracle even change the license for code published as GPL? I am really interested to hear "official" responce from someone that knows this by heart.
On 07/22/2011 01:11 PM, Ljubomir Ljubojevic wrote:
Marian Marinov wrote:
I spoke with my colleges about Ksplice and one very interesting question popped up:
What will happen with our fork efforts if Oracle suddenly patent the Ksplicesoftware ?
I know that in Europe we don't have the problem with software pattents and the GPL will protect us from any issues like that.
However what is the situation in the US and other countries that have software pattents ?
Marian
Any code released as GPL can be forked as long as fork is also GPL. Can Oracle even change the license for code published as GPL?
later code can be published using any license. the license for the earlier code cannot be changed
I am really interested to hear "official" responce from someone that knows this by heart.
Manuel Wolfshant wrote:
later code can be published using any license. the license for the earlier code cannot be changed
So you can take GPL code, change it a little and make it copyrighted (non-GPL)? What about the parts of the code already published under GPL?
On 07/22/2011 12:49 PM, Ljubomir Ljubojevic wrote:
Manuel Wolfshant wrote:
later code can be published using any license. the license for the earlier code cannot be changed
So you can take GPL code, change it a little and make it copyrighted (non-GPL)? What about the parts of the code already published under GPL?
no you can not. gpl does not allow this. once a file is gpl licensed it is so forever until all copyright holders agree to change the license.
grtz
On Fri, 22 Jul 2011, Ljubomir Ljubojevic wrote:
Manuel Wolfshant wrote:
later code can be published using any license. the license for the earlier code cannot be changed
So you can take GPL code, change it a little and make it copyrighted (non-GPL)? What about the parts of the code already published under GPL?
Ljubomir --
Please go participate in Debian-legal, or Fedora-legal or such other non-lawyer debate and speculation societies. This is so clearly out of scope here
-- Russ herrold
On Fri, 22 Jul 2011, Ljubomir Ljubojevic wrote:
Manuel Wolfshant wrote:
later code can be published using any license. the license for the earlier code cannot be changed
So you can take GPL code, change it a little and make it copyrighted (non-GPL)?
It is already copyrighted, whether GPL or not. If you own the copyright, you can distribute a new version which is under a different license, and therefore is not GPL.
If you don't own the copyright, any version you distribute must be GPL.
What about the parts of the code already published under GPL?
Those are already GPL.
Please take the time to read the GPL, and FSF's FAQ on licensing.
Hi,
On 07/22/2011 09:54 AM, Marian Marinov wrote:
I spoke with my colleges about Ksplice and one very interesting question popped up:
What will happen with our fork efforts if Oracle suddenly patent the Ksplicesoftware ?
There are some patents(?) already involved : http://www.faqs.org/patents/assignee/ksplice-inc/
- http://www.faqs.org/patents/app/20100269105 - http://www.faqs.org/patents/app/20100269106
I know that in Europe we don't have the problem with software pattents and the GPL will protect us from any issues like that.
However what is the situation in the US and other countries that have software pattents ?
Carrying content that has patent or regional restrictions is always going to be a tricky question for what we are doing in .centos.org - but there are people we can talk to about stuff like this if there is a need to clear stuff up.
- KB
-
Brandon Ooi -
Charlie Brady -
Cia Watson -
Dmitry E. Mikhailov -
Ferry Huberts -
Gianluca Cecchi -
Hakan Koseoglu -
Jeff Johnson -
Jerry Amundson -
John R. Dennison -
Karanbir Singh -
Khusro Jaleel -
Ljubomir Ljubojevic -
Manuel Wolfshant -
Marian Marinov -
Marko Bevc -
R P Herrold -
Ray Van Dolson -
Sam Wilson -
SZ Quadri -
Tru Huynh -
William Warren