Skip to contentSkip to conversation

nocontrold: Eliminating ocfs2_controld v3

PATCH New Goldwyn Rodrigues 4 participants 6 messages Oct 3, 2013 – Oct 8, 2013 OCFS2

lore
Share & tools
lore
#0 lore raw

cover nocontrold: Eliminating ocfs2_controld v3

Hi,

This is an effort of removing ocfs2_controld.pcmk and getting ocfs2 DLM handling up to the times with respect to DLM (>=4.0.1) and corosync (2.3.x). AFAIK, cman also is being phased out for a unified corosync cluster stack.

fs/dlm performs all the functions with respect to fencing and node management and provides the API's to do so for ocfs2. For all future references, DLM stands for fs/dlm code.

The advantages are:
+ No need to run an additional userspace daemon (ocfs2_controld)
+ No contrrold devince handling and controld protocol
+ Shifting responsibilities of node management to DLM layer

For backward compatibility, we are keeping the controld handling code. Once enough time has passed we can remove a significant portion of the code.

This feature requires modification in the userspace ocfs2-tools. The changes can be found at:
https://github.com/goldwynr/ocfs2-tools branch: nocontrold Currently, not many checks are present in the userspace code, but that would change soon.

These changes were developed on linux-stable 3.11.y, though they are applicable at the current upstream as well. If you want to give the entire kernel a spin, the link is:

https://github.com/goldwynr/linux-stable branch: nocontrold

Changes since v2:
* Joel's comments: patch re-factoring
* No need to safeguard cluster_stack

Changes since v1:
* Backward compatibility with ocfs2_controld
* Moved cluster_stack change restriction to dlm code

Signature · 1 line
Goldwyn
#1 lore raw

cover nocontrold: Eliminating ocfs2_controld v3

On 2013-10-03T00:48:55, Goldwyn Rodrigues <[email protected]> wrote:

Hi Goldwyn,

just a minor comment:

Quoted 2 lines

For backward compatibility, we are keeping the controld handling code. Once enough time has passed we can remove a significant portion of the code.

Can we add this to Documentation/fs/ocfs2.txt, or maybe print a one-time message on activating the deprecated mode with an intended removal date? (I'd propose end-of-2014.)

And, perhaps, a compile-time option to disable this earlier, so that users can choose the stacks they want to build the kernel with.

Thanks,
Lars

Signature · 3 lines
Architect Storage/HA
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imend?rffer, HRB 21284 (AG N?rnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde
#2 lore raw

cover nocontrold: Eliminating ocfs2_controld v3

Quoted 4 lines · Lars Marowsky-Bree

On 2013-10-03T00:48:55, Goldwyn Rodrigues <[email protected]> wrote:

Hi Goldwyn,

Quoted 8 lines

just a minor comment:

Quoted 2 lines

For backward compatibility, we are keeping the controld handling code. Once enough time has passed we can remove a significant portion of the code.

Can we add this to Documentation/fs/ocfs2.txt, or maybe print a one-time message on activating the deprecated mode with an intended removal date? (I'd propose end-of-2014.)

We already print a message in dmesg saying:
printk(KERN_NOTICE "ocfs2: You seem to be using an older "
"version of dlm_controld and/or ocfs2-tools." " Please consider upgrading.\n");

Adding a date to this should not be a problem.

Quoted 4 lines

And, perhaps, a compile-time option to disable this earlier, so that users can choose the stacks they want to build the kernel with.

This would a be a couple of ifdef's in user_cluster_connect(). What do other ocfs2 developers feel about this?

Signature · 1 line
Goldwyn
maintainer
#3 lore raw

cover nocontrold: Eliminating ocfs2_controld v3

Quoted 9 lines · Goldwyn Rodrigues
Quoted 5 lines · Lars Marowsky-Bree

On 2013-10-03T00:48:55, Goldwyn Rodrigues <[email protected]> wrote:

And, perhaps, a compile-time option to disable this earlier, so that users can choose the stacks they want to build the kernel with.

This would a be a couple of ifdef's in user_cluster_connect(). What do other ocfs2 developers feel about this?

We already allow building of o2cb/user stacks via .config. So on that note it isn't terribly different.

lmb, why do you want to enforce non-controld? Shouldn't it be just fine to not install controld and let the filesystem detect a modern enough fs/dlm?

Joel

Quoted 9 lines
Signature · 7 lines
-- 
Goldwyn

_______________________________________________
Ocfs2-devel mailing list
Ocfs2-devel at oss.oracle.com
https://oss.oracle.com/mailman/listinfo/ocfs2-devel
Signature · 6 lines
Life's Little Instruction Book #306

	"Take a nap on Sunday afternoons."

			http://www.jlbec.org/
			jlbec at evilplan.org
#4 lore raw

cover nocontrold: Eliminating ocfs2_controld v3

Quoted 6 lines · Joel Becker

We already allow building of o2cb/user stacks via .config. So on that note it isn't terribly different.

lmb, why do you want to enforce non-controld? Shouldn't it be just fine to not install controld and let the filesystem detect a modern enough fs/dlm?

It's mostly an issue of what the distribution kernels wish to support. One less variable to control for.

We either get that choice upstream or carry an out of tree patch, hence my proposal to do it upstream ;-)

Regards,
Lars

Signature · 3 lines
Architect Storage/HA
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imend?rffer, HRB 21284 (AG N?rnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde
#5 lore raw

cover nocontrold: Eliminating ocfs2_controld v3

In reply to Goldwyn Rodrigues
Quoted 29 lines · On Thu, 3 Oct 2013 00:48:55 -0500 Goldwyn Rodrigues <[email protected]> wrote:

This is an effort of removing ocfs2_controld.pcmk and getting ocfs2 DLM handling up to the times with respect to DLM (>=4.0.1) and corosync (2.3.x). AFAIK, cman also is being phased out for a unified corosync cluster stack.

fs/dlm performs all the functions with respect to fencing and node management and provides the API's to do so for ocfs2. For all future references, DLM stands for fs/dlm code.

The advantages are:
+ No need to run an additional userspace daemon (ocfs2_controld)
+ No contrrold devince handling and controld protocol
+ Shifting responsibilities of node management to DLM layer

For backward compatibility, we are keeping the controld handling code. Once enough time has passed we can remove a significant portion of the code.

This feature requires modification in the userspace ocfs2-tools. The changes can be found at:
https://github.com/goldwynr/ocfs2-tools branch: nocontrold Currently, not many checks are present in the userspace code, but that would change soon.

These changes were developed on linux-stable 3.11.y, though they are applicable at the current upstream as well. If you want to give the entire kernel a spin, the link is:

https://github.com/goldwynr/linux-stable branch: nocontrold

Quoted 7 lines

Changes since v2:
* Joel's comments: patch re-factoring
* No need to safeguard cluster_stack

Changes since v1:
* Backward compatibility with ocfs2_controld
* Moved cluster_stack change restriction to dlm code

I'm not really sure what to do with this patchset. It's too deep for me to merge on my own cognizance so I am dependent upon reviews and acks from ocfs2 specialists before I can move forward. Where are we at with that?