Mailing List Archive

On 2011-07-06T11:47:28, Steven Dake <sdake@redhat.com> wrote:

Hi,

how about not starting a new thread by replying to an existing one? That
would be much nicer for those of us using mail readers that do real
threading ;-)

> Any plans to in some way add ocf semantics into the systemd process
> model? I'd find this would be high value since systemd seems to be the
> direction many distros are headed.

Yes, I think systemd should be powerful enough to replace the LRM. I for
one would welcome that.


Regards,
Lars

--
Architect Storage/HA, OPS Engineering, Novell, Inc.
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde

_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical

On 2011-07-07 12:51, Lars Marowsky-Bree wrote:
> On 2011-07-06T11:47:28, Steven Dake <sdake@redhat.com> wrote:
>
> Hi,
>
> how about not starting a new thread by replying to an existing one? That
> would be much nicer for those of us using mail readers that do real
> threading ;-)
>
>> Any plans to in some way add ocf semantics into the systemd process
>> model? I'd find this would be high value since systemd seems to be the
>> direction many distros are headed.
>
> Yes, I think systemd should be powerful enough to replace the LRM. I for
> one would welcome that.

Which would, of course, be an issue on distros that don't ship systemd
and don't intend to anytime soon. But if lrmd were replaced by a thin
wrapper that encapsulates either systemd or upstart, would that be
workable approach?

Cheers,
Florian

On 07/07/11 21:06, Florian Haas wrote:
> On 2011-07-07 12:51, Lars Marowsky-Bree wrote:
>> On 2011-07-06T11:47:28, Steven Dake<sdake@redhat.com> wrote:
>>
>> Hi,
>>
>> how about not starting a new thread by replying to an existing one? That
>> would be much nicer for those of us using mail readers that do real
>> threading ;-)
>>
>>> Any plans to in some way add ocf semantics into the systemd process
>>> model? I'd find this would be high value since systemd seems to be the
>>> direction many distros are headed.
>>
>> Yes, I think systemd should be powerful enough to replace the LRM. I for
>> one would welcome that.
>
> Which would, of course, be an issue on distros that don't ship systemd
> and don't intend to anytime soon. But if lrmd were replaced by a thin
> wrapper that encapsulates either systemd or upstart, would that be
> workable approach?

IIRC Beekhof said something about Matahari being a potential LRM
replacement. Can she work as our thin wrapper in this context, or am I
getting confused?

Regards,

Tim
--
Tim Serong <tserong@novell.com>
Senior Clustering Engineer, OPS Engineering, Novell Inc.
_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical

On Thu, Jul 07, 2011 at 10:56:18PM +1000, Tim Serong wrote:
> On 07/07/11 21:06, Florian Haas wrote:
> > On 2011-07-07 12:51, Lars Marowsky-Bree wrote:
> >> On 2011-07-06T11:47:28, Steven Dake<sdake@redhat.com> wrote:
> >>
> >> Hi,
> >>
> >> how about not starting a new thread by replying to an existing one? That
> >> would be much nicer for those of us using mail readers that do real
> >> threading ;-)
> >>
> >>> Any plans to in some way add ocf semantics into the systemd process
> >>> model? I'd find this would be high value since systemd seems to be the
> >>> direction many distros are headed.
> >>
> >> Yes, I think systemd should be powerful enough to replace the LRM. I for
> >> one would welcome that.

That's news for me. But if so, I'd certainly be happy too.

> > Which would, of course, be an issue on distros that don't ship systemd
> > and don't intend to anytime soon. But if lrmd were replaced by a thin
> > wrapper that encapsulates either systemd or upstart, would that be
> > workable approach?
>
> IIRC Beekhof said something about Matahari being a potential LRM
> replacement. Can she work as our thin wrapper in this context, or am I
> getting confused?

Don't know anything about Matahari, but I guess it's something
brand new. If so, why then have that developed if it could be
replaced by systemd? /me confused too :-)

Cheers,

Dejan

> Regards,
>
> Tim
> --
> Tim Serong <tserong@novell.com>
> Senior Clustering Engineer, OPS Engineering, Novell Inc.
> _______________________________________________
> ha-wg-technical mailing list
> ha-wg-technical@lists.linux-foundation.org
> https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical
_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical

On 07/07/2011 07:09 AM, Dejan Muhamedagic wrote:
> On Thu, Jul 07, 2011 at 10:56:18PM +1000, Tim Serong wrote:
>> On 07/07/11 21:06, Florian Haas wrote:
>>> On 2011-07-07 12:51, Lars Marowsky-Bree wrote:
>>>> On 2011-07-06T11:47:28, Steven Dake<sdake@redhat.com> wrote:
>>>>
>>>> Hi,
>>>>
>>>> how about not starting a new thread by replying to an existing one? That
>>>> would be much nicer for those of us using mail readers that do real
>>>> threading ;-)
>>>>
>>>>> Any plans to in some way add ocf semantics into the systemd process
>>>>> model? I'd find this would be high value since systemd seems to be the
>>>>> direction many distros are headed.
>>>>
>>>> Yes, I think systemd should be powerful enough to replace the LRM. I for
>>>> one would welcome that.
>
> That's news for me. But if so, I'd certainly be happy too.
>
>>> Which would, of course, be an issue on distros that don't ship systemd
>>> and don't intend to anytime soon. But if lrmd were replaced by a thin
>>> wrapper that encapsulates either systemd or upstart, would that be
>>> workable approach?
>>
>> IIRC Beekhof said something about Matahari being a potential LRM
>> replacement. Can she work as our thin wrapper in this context, or am I
>> getting confused?
>
> Don't know anything about Matahari, but I guess it's something
> brand new. If so, why then have that developed if it could be
> replaced by systemd? /me confused too :-)
>

matahari includes a "service" daemon which provides the various models
needed. This matahari-serviced could be modified to work with systemd
and upstart directly rather then through service scripts.

Regards
-steve
> Cheers,
>
> Dejan
>
>> Regards,
>>
>> Tim
>> --
>> Tim Serong <tserong@novell.com>
>> Senior Clustering Engineer, OPS Engineering, Novell Inc.
>> _______________________________________________
>> ha-wg-technical mailing list
>> ha-wg-technical@lists.linux-foundation.org
>> https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical
> _______________________________________________
> ha-wg-technical mailing list
> ha-wg-technical@lists.linux-foundation.org
> https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical

_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical

On Fri, Jul 8, 2011 at 2:17 AM, Steven Dake <sdake@redhat.com> wrote:
> On 07/07/2011 07:09 AM, Dejan Muhamedagic wrote:
>> On Thu, Jul 07, 2011 at 10:56:18PM +1000, Tim Serong wrote:
>>> On 07/07/11 21:06, Florian Haas wrote:
>>>> On 2011-07-07 12:51, Lars Marowsky-Bree wrote:
>>>>> On 2011-07-06T11:47:28, Steven Dake<sdake@redhat.com>  wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> how about not starting a new thread by replying to an existing one? That
>>>>> would be much nicer for those of us using mail readers that do real
>>>>> threading ;-)
>>>>>
>>>>>> Any plans to in some way add ocf semantics into the systemd process
>>>>>> model?  I'd find this would be high value since systemd seems to be the
>>>>>> direction many distros are headed.
>>>>>
>>>>> Yes, I think systemd should be powerful enough to replace the LRM. I for
>>>>> one would welcome that.
>>
>> That's news for me. But if so, I'd certainly be happy too.
>>
>>>> Which would, of course, be an issue on distros that don't ship systemd
>>>> and don't intend to anytime soon. But if lrmd were replaced by a thin
>>>> wrapper that encapsulates either systemd or upstart, would that be
>>>> workable approach?
>>>
>>> IIRC Beekhof said something about Matahari being a potential LRM
>>> replacement.

Correct

>>> Can she work as our thin wrapper in this context, or am I
>>> getting confused?

Not really, it would be the thing getting wrapped.

>>
>> Don't know anything about Matahari, but I guess it's something
>> brand new.

I gave a talk about it at Plumbers

>> If so, why then have that developed if it could be
>> replaced by systemd?

It cannot be.
Even if systemd started supporting OCF-style parameters (there is no
indication they're interested in this), it does not run on windows and
is more heavy weight than we need.

>> /me confused too :-)
>>
>
> matahari includes a "service" daemon which provides the various models
> needed.  This matahari-serviced could be modified to work with systemd
> and upstart directly rather then through service scripts.

Right. This is what will probably end up happening.
Sooner rather than later too given the quality of their LSB compatibility.
_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical

On 2011-07-07T16:09:55, Dejan Muhamedagic <dejan@suse.de> wrote:

> > >>> Any plans to in some way add ocf semantics into the systemd process
> > >>> model? I'd find this would be high value since systemd seems to be the
> > >>> direction many distros are headed.
> > >> Yes, I think systemd should be powerful enough to replace the LRM. I for
> > >> one would welcome that.
> That's news for me. But if so, I'd certainly be happy too.

Well, yes, it is news, because this is the first time it was discussed.
;-)

If we could hook into the basic system-wide "init" instead of using our
own, that would clearly be desirable.

(Will respond to the wrapper issue in another mail.)

Regards,
Lars

--
Architect Storage/HA, OPS Engineering, Novell, Inc.
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde

_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical

On 2011-07-08T10:33:58, Andrew Beekhof <andrew@beekhof.net> wrote:

> >>> Can she work as our thin wrapper in this context, or am I
> >>> getting confused?
> Not really, it would be the thing getting wrapped.

Okay, now I'm a bit confused too.

So we'd end up wrapping matahari, systemd, lrmd (for legacy systems),
possibly upstart too?

I foresee a wrapper that is more complex than continuing to provide our
own implementation in the first place.


Regards,
Lars

--
Architect Storage/HA, OPS Engineering, Novell, Inc.
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde

_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical

On 07/08/2011 10:19 AM, Lars Marowsky-Bree wrote:
> On 2011-07-08T10:33:58, Andrew Beekhof <andrew@beekhof.net> wrote:
>
>>>>> Can she work as our thin wrapper in this context, or am I
>>>>> getting confused?
>> Not really, it would be the thing getting wrapped.
>
> Okay, now I'm a bit confused too.
>
> So we'd end up wrapping matahari, systemd, lrmd (for legacy systems),
> possibly upstart too?
>
> I foresee a wrapper that is more complex than continuing to provide our
> own implementation in the first place.

So, perhaps we would consider adding systemd support to lrmd instead,
like has already happened for upstart?

Florian

On Fri, Jul 8, 2011 at 6:19 PM, Lars Marowsky-Bree <lmb@suse.de> wrote:
> On 2011-07-08T10:33:58, Andrew Beekhof <andrew@beekhof.net> wrote:
>
>> >>> Can she work as our thin wrapper in this context, or am I
>> >>> getting confused?
>> Not really, it would be the thing getting wrapped.
>
> Okay, now I'm a bit confused too.
>
> So we'd end up wrapping matahari, systemd, lrmd (for legacy systems),
> possibly upstart too?

No.

lrmd == matahari system agent lib + wrapper, or
lrmd == matahari system agent via qmf or dbus interfaces

The libs/agent would add systemd and upstart support to the existing
lsb and ocf.
Much like lrmd would but without the plugins.

>
> I foresee a wrapper that is more complex than continuing to provide our
> own implementation in the first place.
>
>
> Regards,
>    Lars
>
> --
> Architect Storage/HA, OPS Engineering, Novell, Inc.
> SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nürnberg)
> "Experience is the name everyone gives to their mistakes." -- Oscar Wilde
>
>
_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical