[opensuse-factory] Tumbleweed inconsistency between DRBD kernel module and tools
Hi all, yesterday I startet to migrate my openSUSE machines from normal distribution to the rolling-update Tumbleweed (currently openSUSE 12.3) - machines are productive and so Factory was no option. After switching to Tumbleweed everything works fine except DRBD. I found out that there is an inconsistency between drbd kernel module (V8.4.3 in kernel-default) and the drbd admin tools (V8.3.11 from repository) which ends up with the inability to start the cluster volumes. As workaround I found a matching rpm on suse build service, but please include a proper drbd-tools version in the Tumbleweed repos. Thx in advance. Manfred -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/08/2013 04:01 AM, Wirlach Manfred pecked at the keyboard and wrote:
Hi all,
yesterday I startet to migrate my openSUSE machines from normal distribution to the rolling-update Tumbleweed (currently openSUSE 12.3) - machines are productive and so Factory was no option.
After switching to Tumbleweed everything works fine except DRBD. I found out that there is an inconsistency between drbd kernel module (V8.4.3 in kernel-default) and the drbd admin tools (V8.3.11 from repository) which ends up with the inability to start the cluster volumes.
As workaround I found a matching rpm on suse build service, but please include a proper drbd-tools version in the Tumbleweed repos.
Thx in advance.
Manfred
Stability, stability, stability is what is needed for production machines which is something that TW _cannot_ provide/guarantee. Therefore it is not recommended for your use case. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Tue, Oct 08, 2013 at 12:07:18PM -0400, Ken Schneider - openSUSE wrote:
On 10/08/2013 04:01 AM, Wirlach Manfred pecked at the keyboard and wrote:
Hi all,
yesterday I startet to migrate my openSUSE machines from normal distribution to the rolling-update Tumbleweed (currently openSUSE 12.3) - machines are productive and so Factory was no option.
After switching to Tumbleweed everything works fine except DRBD. I found out that there is an inconsistency between drbd kernel module (V8.4.3 in kernel-default) and the drbd admin tools (V8.3.11 from repository) which ends up with the inability to start the cluster volumes.
As workaround I found a matching rpm on suse build service, but please include a proper drbd-tools version in the Tumbleweed repos.
Thx in advance.
Manfred
Stability, stability, stability is what is needed for production machines which is something that TW _cannot_ provide/guarantee. Therefore it is not recommended for your use case.
It all depends on what your "production" is? I know lots of valid use cases for constantly rolling updates for real-life-production-systems, so please don't think that "stability" is somehow the only way stuff like that can work. I'll be glad to add the drbd tools to Tumbleweed, Wirlach, what repo should I pull them from to ensure they are kept up to date with the kernel releases? thanks, greg k-h -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/09/2013 12:21 AM, Greg KH pecked at the keyboard and wrote:
On Tue, Oct 08, 2013 at 12:07:18PM -0400, Ken Schneider - openSUSE wrote:
On 10/08/2013 04:01 AM, Wirlach Manfred pecked at the keyboard and wrote:
Hi all,
yesterday I startet to migrate my openSUSE machines from normal distribution to the rolling-update Tumbleweed (currently openSUSE 12.3) - machines are productive and so Factory was no option.
After switching to Tumbleweed everything works fine except DRBD. I found out that there is an inconsistency between drbd kernel module (V8.4.3 in kernel-default) and the drbd admin tools (V8.3.11 from repository) which ends up with the inability to start the cluster volumes.
As workaround I found a matching rpm on suse build service, but please include a proper drbd-tools version in the Tumbleweed repos.
Thx in advance.
Manfred
Stability, stability, stability is what is needed for production machines which is something that TW _cannot_ provide/guarantee. Therefore it is not recommended for your use case.
It all depends on what your "production" is? I know lots of valid use cases for constantly rolling updates for real-life-production-systems, so please don't think that "stability" is somehow the only way stuff like that can work.
I'll be glad to add the drbd tools to Tumbleweed, Wirlach, what repo should I pull them from to ensure they are kept up to date with the kernel releases?
thanks,
greg k-h
Properly administered/managed /production/ machines should never be updated without properly testing any software update whether it be kernel or application updates. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/09/2013 12:44 PM, Ken Schneider - openSUSE pecked at the keyboard and wrote:
On 10/09/2013 12:21 AM, Greg KH pecked at the keyboard and wrote:
On Tue, Oct 08, 2013 at 12:07:18PM -0400, Ken Schneider - openSUSE wrote:
On 10/08/2013 04:01 AM, Wirlach Manfred pecked at the keyboard and wrote:
Hi all,
yesterday I startet to migrate my openSUSE machines from normal distribution to the rolling-update Tumbleweed (currently openSUSE 12.3) - machines are productive and so Factory was no option.
After switching to Tumbleweed everything works fine except DRBD. I found out that there is an inconsistency between drbd kernel module (V8.4.3 in kernel-default) and the drbd admin tools (V8.3.11 from repository) which ends up with the inability to start the cluster volumes.
As workaround I found a matching rpm on suse build service, but please include a proper drbd-tools version in the Tumbleweed repos.
Thx in advance.
Manfred
Stability, stability, stability is what is needed for production machines which is something that TW _cannot_ provide/guarantee. Therefore it is not recommended for your use case.
It all depends on what your "production" is? I know lots of valid use cases for constantly rolling updates for real-life-production-systems, so please don't think that "stability" is somehow the only way stuff like that can work.
I'll be glad to add the drbd tools to Tumbleweed, Wirlach, what repo should I pull them from to ensure they are kept up to date with the kernel releases?
thanks,
greg k-h
Properly administered/managed /production/ machines should never be updated without properly testing any software update whether it be kernel or application updates.
Tested on a test machine of course. The IT direstor at the places I worked at would fire anyone for allowing -untested- updates to production machines. I don't want to imply that TW is not stable just that bugs will always exist the the least likely place. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/09/2013 12:44 PM, Ken Schneider - openSUSE pecked at the keyboard and wrote:
On 10/09/2013 12:21 AM, Greg KH pecked at the keyboard and wrote:
On Tue, Oct 08, 2013 at 12:07:18PM -0400, Ken Schneider - openSUSE wrote:
On 10/08/2013 04:01 AM, Wirlach Manfred pecked at the keyboard and wrote:
Hi all,
yesterday I startet to migrate my openSUSE machines from normal distribution to the rolling-update Tumbleweed (currently openSUSE 12.3) - machines are productive and so Factory was no option.
After switching to Tumbleweed everything works fine except DRBD. I found out that there is an inconsistency between drbd kernel module (V8.4.3 in kernel-default) and the drbd admin tools (V8.3.11 from repository) which ends up with the inability to start the cluster volumes.
As workaround I found a matching rpm on suse build service, but please include a proper drbd-tools version in the Tumbleweed repos.
Thx in advance.
Manfred
Stability, stability, stability is what is needed for production machines which is something that TW _cannot_ provide/guarantee. Therefore it is not recommended for your use case.
It all depends on what your "production" is? I know lots of valid use cases for constantly rolling updates for real-life-production-systems, so please don't think that "stability" is somehow the only way stuff like that can work.
I'll be glad to add the drbd tools to Tumbleweed, Wirlach, what repo should I pull them from to ensure they are kept up to date with the kernel releases?
thanks,
greg k-h
Properly administered/managed /production/ machines should never be updated without properly testing any software update whether it be kernel or application updates.
Tested on a test machine of course. The IT direstor at the places I worked at would fire anyone for allowing -untested- updates to production machines.
I don't want to imply that TW is not stable just that bugs will always exist the the least likely place.
-- Ken Schneider SuSe since Version 5.2, June 1998 -- Every serious IT admin should test its software before deployment, but hey - no risk, no fun;-) I always try to do it, but be honest - do you really test ALL services on the machines? How large is you test environment? How do you test apps with complex dependencies from other services?
My experience is that we all test the main applications well, but many problems come along with commodity services in background and slightly changed default parameters or permissions. For instance the Tumbleweed update changes the hostname to its FQDN. The apps worked well after upgrade (opensuse guys did a great job!), but corosync cluster and drbd had problems due to changed node names in cluster config... Best Regards Manfred -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/09/13 18:44, Ken Schneider - openSUSE wrote:
On 10/09/2013 12:21 AM, Greg KH pecked at the keyboard and wrote:
It all depends on what your "production" is? I know lots of valid use cases for constantly rolling updates for real-life-production-systems, so please don't think that "stability" is somehow the only way stuff like that can work.
Properly administered/managed /production/ machines should never be updated without properly testing any software update whether it be kernel or application updates.
Google, Netflix, Facebook, and other companies beg to differ. Netflix goes as far as openly killing production machines in the wild. http://techblog.netflix.com/2012/07/chaos-monkey-released-into-wild.html is the best link I could find within 5 seconds. Facebook DevOps' releases dozens of production updates directly from developer's trunk each day. If you have a fault-tolerant distributed large scale system, "stability" needs a new definition; and thoroughly testing each and every change doesn't belong to that definition. It defies other goals of flexibility and ability to change very quick. The goal is that deployment / installation / updates is not something special that must be planned for. It's something that happens all the time, around the clock, in quick and rapid steps. If you want to stay current, you have to look at what these guys are doing. You have to evaluate where these approaches fit in your environment and where they don't. (Not that they're not testing; continous integration is an obvious part of their dev and ops process. It's how they are testing and how they are coping with failures that makes the difference.) If you just oppose that change, well, look up Gorbatchow's famous last words (which he never said, btw) »those who are too late are comdemned by life«. If you work as a sysadmin, have a look at this. Otherwise, in one to two decades, you won't have work any more. Joachim -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Joachim Schrod, Roedermark, Germany Email: jschrod@acm.org -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/09/2013 08:16 PM, Joachim Schrod pecked at the keyboard and wrote:
On 10/09/13 18:44, Ken Schneider - openSUSE wrote:
On 10/09/2013 12:21 AM, Greg KH pecked at the keyboard and wrote:
It all depends on what your "production" is? I know lots of valid use cases for constantly rolling updates for real-life-production-systems, so please don't think that "stability" is somehow the only way stuff like that can work.
Properly administered/managed /production/ machines should never be updated without properly testing any software update whether it be kernel or application updates.
Google, Netflix, Facebook, and other companies beg to differ.
They don't use single purpose machines and can probably take 100's of machines offline and still function. Try comparing apples to apples. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/10/13 05:35, Ken Schneider - openSUSE wrote:
On 10/09/2013 08:16 PM, Joachim Schrod pecked at the keyboard and wrote:
On 10/09/13 18:44, Ken Schneider - openSUSE wrote:
On 10/09/2013 12:21 AM, Greg KH pecked at the keyboard and wrote:
It all depends on what your "production" is? I know lots of valid use cases for constantly rolling updates for real-life-production-systems, so please don't think that "stability" is somehow the only way stuff like that can work.
Properly administered/managed /production/ machines should never be updated without properly testing any software update whether it be kernel or application updates.
Google, Netflix, Facebook, and other companies beg to differ.
They don't use single purpose machines and can probably take 100's of machines offline and still function. Try comparing apples to apples.
So, can you explain to me, why these are not `Properly administered/managed/production/ machines' (your words), just because there are lots of them? As Greg expressed it so clearly: `It all depends on what your "production" is'. Where you replied with the term *never*, without looking at such use cases that get increasingly more common. Joachim -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Joachim Schrod, Roedermark, Germany Email: jschrod@acm.org -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/10/2013 12:46 PM, Joachim Schrod pecked at the keyboard and wrote:
On 10/10/13 05:35, Ken Schneider - openSUSE wrote:
On 10/09/2013 08:16 PM, Joachim Schrod pecked at the keyboard and wrote:
On 10/09/13 18:44, Ken Schneider - openSUSE wrote:
On 10/09/2013 12:21 AM, Greg KH pecked at the keyboard and wrote:
It all depends on what your "production" is? I know lots of valid use cases for constantly rolling updates for real-life-production-systems, so please don't think that "stability" is somehow the only way stuff like that can work.
Properly administered/managed /production/ machines should never be updated without properly testing any software update whether it be kernel or application updates.
Google, Netflix, Facebook, and other companies beg to differ.
They don't use single purpose machines and can probably take 100's of machines offline and still function. Try comparing apples to apples.
So, can you explain to me, why these are not `Properly administered/managed/production/ machines' (your words), just because there are lots of them?
As Greg expressed it so clearly: `It all depends on what your "production" is'. Where you replied with the term *never*, without looking at such use cases that get increasingly more common.
Joachim
It may be becoming the norm/common in LARGE enterprises, but all of the smaller mom and pop places cannot afford a server farm (the majority of the use case here). Enough said. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 El 2013-10-08 a las 08:01 -0000, Wirlach Manfred escribió:
yesterday I startet to migrate my openSUSE machines from normal distribution to the rolling-update Tumbleweed (currently openSUSE 12.3) - machines are productive and so Factory was no option.
Neither is tumbleweed... (IMNSHO) - -- Cheers Carlos E. R. (from 11.4, with Evergreen, x86_64 "Celadon" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) iF4EAREIAAYFAlJUltEACgkQja8UbcUWM1zJygD8C45wenPVvCh00XrKVrYGiHOH MFGe/xBHG+lFtkRr1i4A+wb2CIGfzync7RhzQNFLqJ0nxydgd7hNjDgphG9YiwSg =oWgc -----END PGP SIGNATURE-----
participants (5)
-
Carlos E. R. -
Greg KH -
Joachim Schrod -
Ken Schneider - openSUSE -
Wirlach Manfred