Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Linux-HA: Dev

LVM monitor change

 

 

Linux-HA dev RSS feed   Index | Next | Previous | View Threaded


dejan at suse

Apr 4, 2012, 9:29 AM

Post #1 of 11 (977 views)
Permalink
LVM monitor change

Hi all,

This is a proposed set of two patches which would eliminate use
of LVM commands in the monitor path. We already discussed the
issue elsewhere and I don't see any point in keeping
vgck/vgdisplay given that they don't result in better monitoring
under normal circumstances. And if the circumstances are such
that the new monitoring fails, I think that there'll be many
more problems on the node than a failed volume group.

Cheers,

Dejan
Attachments: LVM.vgck.patch (2.54 KB)


renayama19661014 at ybb

Apr 4, 2012, 7:32 PM

Post #2 of 11 (943 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Dejan,

I agree to your patch.

Best Regards,
Hideo Yamauchi.

--- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:

> Hi all,
>
> This is a proposed set of two patches which would eliminate use
> of LVM commands in the monitor path. We already discussed the
> issue elsewhere and I don't see any point in keeping
> vgck/vgdisplay given that they don't result in better monitoring
> under normal circumstances. And if the circumstances are such
> that the new monitoring fails, I think that there'll be many
> more problems on the node than a failed volume group.
>
> Cheers,
>
> Dejan
>
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/


dejan at suse

Apr 5, 2012, 8:37 AM

Post #3 of 11 (946 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Hideo-san,

On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> Hi Dejan,
>
> I agree to your patch.

Thank you for the reply.

BTW, the monitor was shamelessly stolen from Vladislav.

Applied.

ocft test passed (after some struggle and eventually fixing the
ocft source).

Cheers,

Dejan

> Best Regards,
> Hideo Yamauchi.
>
> --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
>
> > Hi all,
> >
> > This is a proposed set of two patches which would eliminate use
> > of LVM commands in the monitor path. We already discussed the
> > issue elsewhere and I don't see any point in keeping
> > vgck/vgdisplay given that they don't result in better monitoring
> > under normal circumstances. And if the circumstances are such
> > that the new monitoring fails, I think that there'll be many
> > more problems on the node than a failed volume group.
> >
> > Cheers,
> >
> > Dejan
> >
> _______________________________________________________
> Linux-HA-Dev: Linux-HA-Dev [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/


renayama19661014 at ybb

Apr 5, 2012, 6:50 PM

Post #4 of 11 (941 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Dejan,

I change validate-all and want to change it to always carry out validate-all.
I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.

How do you think?

Best Regards,
Hideo Yamauchi.

--- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:

> Hi Hideo-san,
>
> On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> > Hi Dejan,
> >
> > I agree to your patch.
>
> Thank you for the reply.
>
> BTW, the monitor was shamelessly stolen from Vladislav.
>
> Applied.
>
> ocft test passed (after some struggle and eventually fixing the
> ocft source).
>
> Cheers,
>
> Dejan
>
> > Best Regards,
> > Hideo Yamauchi.
> >
> > --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
> >
> > > Hi all,
> > >
> > > This is a proposed set of two patches which would eliminate use
> > > of LVM commands in the monitor path. We already discussed the
> > > issue elsewhere and I don't see any point in keeping
> > > vgck/vgdisplay given that they don't result in better monitoring
> > > under normal circumstances. And if the circumstances are such
> > > that the new monitoring fails, I think that there'll be many
> > > more problems on the node than a failed volume group.
> > >
> > > Cheers,
> > >
> > > Dejan
> > >
> > _______________________________________________________
> > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > Home Page: http://linux-ha.org/
> _______________________________________________________
> Linux-HA-Dev: Linux-HA-Dev [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
>
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/


dejan at suse

Apr 6, 2012, 4:00 AM

Post #5 of 11 (939 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Hideo-san,

On Fri, Apr 06, 2012 at 10:50:39AM +0900, renayama19661014 [at] ybb wrote:
> Hi Dejan,
>
> I change validate-all and want to change it to always carry out validate-all.
> I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
>
> How do you think?

Isn't it that validate-all may be really necessary only in the
start action? The repeating monitor is scheduled only after a
successful start.

Cheers,

Dejan

> Best Regards,
> Hideo Yamauchi.
>
> --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
>
> > Hi Hideo-san,
> >
> > On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> > > Hi Dejan,
> > >
> > > I agree to your patch.
> >
> > Thank you for the reply.
> >
> > BTW, the monitor was shamelessly stolen from Vladislav.
> >
> > Applied.
> >
> > ocft test passed (after some struggle and eventually fixing the
> > ocft source).
> >
> > Cheers,
> >
> > Dejan
> >
> > > Best Regards,
> > > Hideo Yamauchi.
> > >
> > > --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
> > >
> > > > Hi all,
> > > >
> > > > This is a proposed set of two patches which would eliminate use
> > > > of LVM commands in the monitor path. We already discussed the
> > > > issue elsewhere and I don't see any point in keeping
> > > > vgck/vgdisplay given that they don't result in better monitoring
> > > > under normal circumstances. And if the circumstances are such
> > > > that the new monitoring fails, I think that there'll be many
> > > > more problems on the node than a failed volume group.
> > > >
> > > > Cheers,
> > > >
> > > > Dejan
> > > >
> > > _______________________________________________________
> > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > Home Page: http://linux-ha.org/
> > _______________________________________________________
> > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > Home Page: http://linux-ha.org/
> >
> _______________________________________________________
> Linux-HA-Dev: Linux-HA-Dev [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/


renayama19661014 at ybb

Apr 8, 2012, 5:18 PM

Post #6 of 11 (936 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Dejan,

Thank you for comments.

> > I change validate-all and want to change it to always carry out validate-all.
> > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> >
> > How do you think?
>
> Isn't it that validate-all may be really necessary only in the
> start action? The repeating monitor is scheduled only after a
> successful start.

It may be surely necessary as you say.
However, I think validate-all to unify it so that it is always carried out.

How about what the check of vgck/vgdisplay chooses it in a parameter and can carry out?


Best Regards,
Hideo Yamauchi.

--- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:

> Hi Hideo-san,
>
> On Fri, Apr 06, 2012 at 10:50:39AM +0900, renayama19661014 [at] ybb wrote:
> > Hi Dejan,
> >
> > I change validate-all and want to change it to always carry out validate-all.
> > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> >
> > How do you think?
>
> Isn't it that validate-all may be really necessary only in the
> start action? The repeating monitor is scheduled only after a
> successful start.
>
> Cheers,
>
> Dejan
>
> > Best Regards,
> > Hideo Yamauchi.
> >
> > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> >
> > > Hi Hideo-san,
> > >
> > > On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> > > > Hi Dejan,
> > > >
> > > > I agree to your patch.
> > >
> > > Thank you for the reply.
> > >
> > > BTW, the monitor was shamelessly stolen from Vladislav.
> > >
> > > Applied.
> > >
> > > ocft test passed (after some struggle and eventually fixing the
> > > ocft source).
> > >
> > > Cheers,
> > >
> > > Dejan
> > >
> > > > Best Regards,
> > > > Hideo Yamauchi.
> > > >
> > > > --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > >
> > > > > Hi all,
> > > > >
> > > > > This is a proposed set of two patches which would eliminate use
> > > > > of LVM commands in the monitor path. We already discussed the
> > > > > issue elsewhere and I don't see any point in keeping
> > > > > vgck/vgdisplay given that they don't result in better monitoring
> > > > > under normal circumstances. And if the circumstances are such
> > > > > that the new monitoring fails, I think that there'll be many
> > > > > more problems on the node than a failed volume group.
> > > > >
> > > > > Cheers,
> > > > >
> > > > > Dejan
> > > > >
> > > > _______________________________________________________
> > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > Home Page: http://linux-ha.org/
> > > _______________________________________________________
> > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > Home Page: http://linux-ha.org/
> > >
> > _______________________________________________________
> > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > Home Page: http://linux-ha.org/
>
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/


renayama19661014 at ybb

Apr 8, 2012, 7:07 PM

Post #7 of 11 (938 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Dejan,

I send the patch which I thought about just to make sure.

Change 1) Changed validate-all to the parameter checks. And always call it.
Change 2) Chose vgck/vgdisplay by start processing and carried it out.

Best Regards,
Hideo Yamauchi.

--- On Mon, 2012/4/9, renayama19661014 [at] ybb <renayama19661014 [at] ybb> wrote:

> Hi Dejan,
>
> Thank you for comments.
>
> > > I change validate-all and want to change it to always carry out validate-all.
> > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > >
> > > How do you think?
> >
> > Isn't it that validate-all may be really necessary only in the
> > start action? The repeating monitor is scheduled only after a
> > successful start.
>
> It may be surely necessary as you say.
> However, I think validate-all to unify it so that it is always carried out.
>
> How about what the check of vgck/vgdisplay chooses it in a parameter and can carry out?
>
>
> Best Regards,
> Hideo Yamauchi.
>
> --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
>
> > Hi Hideo-san,
> >
> > On Fri, Apr 06, 2012 at 10:50:39AM +0900, renayama19661014 [at] ybb wrote:
> > > Hi Dejan,
> > >
> > > I change validate-all and want to change it to always carry out validate-all.
> > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > >
> > > How do you think?
> >
> > Isn't it that validate-all may be really necessary only in the
> > start action? The repeating monitor is scheduled only after a
> > successful start.
> >
> > Cheers,
> >
> > Dejan
> >
> > > Best Regards,
> > > Hideo Yamauchi.
> > >
> > > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> > >
> > > > Hi Hideo-san,
> > > >
> > > > On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> > > > > Hi Dejan,
> > > > >
> > > > > I agree to your patch.
> > > >
> > > > Thank you for the reply.
> > > >
> > > > BTW, the monitor was shamelessly stolen from Vladislav.
> > > >
> > > > Applied.
> > > >
> > > > ocft test passed (after some struggle and eventually fixing the
> > > > ocft source).
> > > >
> > > > Cheers,
> > > >
> > > > Dejan
> > > >
> > > > > Best Regards,
> > > > > Hideo Yamauchi.
> > > > >
> > > > > --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > > >
> > > > > > Hi all,
> > > > > >
> > > > > > This is a proposed set of two patches which would eliminate use
> > > > > > of LVM commands in the monitor path. We already discussed the
> > > > > > issue elsewhere and I don't see any point in keeping
> > > > > > vgck/vgdisplay given that they don't result in better monitoring
> > > > > > under normal circumstances. And if the circumstances are such
> > > > > > that the new monitoring fails, I think that there'll be many
> > > > > > more problems on the node than a failed volume group.
> > > > > >
> > > > > > Cheers,
> > > > > >
> > > > > > Dejan
> > > > > >
> > > > > _______________________________________________________
> > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > Home Page: http://linux-ha.org/
> > > > _______________________________________________________
> > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > Home Page: http://linux-ha.org/
> > > >
> > > _______________________________________________________
> > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > Home Page: http://linux-ha.org/
> >
> _______________________________________________________
> Linux-HA-Dev: Linux-HA-Dev [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
>
Attachments: LVM.patch.0410 (3.92 KB)


dejan at suse

Apr 9, 2012, 8:59 AM

Post #8 of 11 (942 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Hideo-san,

On Mon, Apr 09, 2012 at 09:18:07AM +0900, renayama19661014 [at] ybb wrote:
> Hi Dejan,
>
> Thank you for comments.
>
> > > I change validate-all and want to change it to always carry out validate-all.
> > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > >
> > > How do you think?
> >
> > Isn't it that validate-all may be really necessary only in the
> > start action? The repeating monitor is scheduled only after a
> > successful start.
>
> It may be surely necessary as you say.
> However, I think validate-all to unify it so that it is always carried out.

But why?

> How about what the check of vgck/vgdisplay chooses it in a parameter and can carry out?

Again, why? It doesn't make any difference for a running
resource? We may do this before the start operation, of course.

Cheers,

Dejan

>
> Best Regards,
> Hideo Yamauchi.
>
> --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
>
> > Hi Hideo-san,
> >
> > On Fri, Apr 06, 2012 at 10:50:39AM +0900, renayama19661014 [at] ybb wrote:
> > > Hi Dejan,
> > >
> > > I change validate-all and want to change it to always carry out validate-all.
> > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > >
> > > How do you think?
> >
> > Isn't it that validate-all may be really necessary only in the
> > start action? The repeating monitor is scheduled only after a
> > successful start.
> >
> > Cheers,
> >
> > Dejan
> >
> > > Best Regards,
> > > Hideo Yamauchi.
> > >
> > > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> > >
> > > > Hi Hideo-san,
> > > >
> > > > On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> > > > > Hi Dejan,
> > > > >
> > > > > I agree to your patch.
> > > >
> > > > Thank you for the reply.
> > > >
> > > > BTW, the monitor was shamelessly stolen from Vladislav.
> > > >
> > > > Applied.
> > > >
> > > > ocft test passed (after some struggle and eventually fixing the
> > > > ocft source).
> > > >
> > > > Cheers,
> > > >
> > > > Dejan
> > > >
> > > > > Best Regards,
> > > > > Hideo Yamauchi.
> > > > >
> > > > > --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > > >
> > > > > > Hi all,
> > > > > >
> > > > > > This is a proposed set of two patches which would eliminate use
> > > > > > of LVM commands in the monitor path. We already discussed the
> > > > > > issue elsewhere and I don't see any point in keeping
> > > > > > vgck/vgdisplay given that they don't result in better monitoring
> > > > > > under normal circumstances. And if the circumstances are such
> > > > > > that the new monitoring fails, I think that there'll be many
> > > > > > more problems on the node than a failed volume group.
> > > > > >
> > > > > > Cheers,
> > > > > >
> > > > > > Dejan
> > > > > >
> > > > > _______________________________________________________
> > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > Home Page: http://linux-ha.org/
> > > > _______________________________________________________
> > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > Home Page: http://linux-ha.org/
> > > >
> > > _______________________________________________________
> > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > Home Page: http://linux-ha.org/
> >
> _______________________________________________________
> Linux-HA-Dev: Linux-HA-Dev [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/


renayama19661014 at ybb

Apr 9, 2012, 8:43 PM

Post #9 of 11 (929 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Dejan,

Thank you for comments.


> Hi Hideo-san,
>
> On Mon, Apr 09, 2012 at 09:18:07AM +0900, renayama19661014 [at] ybb wrote:
> > Hi Dejan,
> >
> > Thank you for comments.
> >
> > > > I change validate-all and want to change it to always carry out validate-all.
> > > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > > >
> > > > How do you think?
> > >
> > > Isn't it that validate-all may be really necessary only in the
> > > start action? The repeating monitor is scheduled only after a
> > > successful start.
> >
> > It may be surely necessary as you say.
> > However, I think validate-all to unify it so that it is always carried out.
>
> But why?

There is the resource to carry out validate-all every time a lot.
We wish it becomes LVM in the same way.

>
> > How about what the check of vgck/vgdisplay chooses it in a parameter and can carry out?
>
> Again, why? It doesn't make any difference for a running
> resource? We may do this before the start operation, of course.

My correction is different from original LVM in big validate-all.

There were many mistakes to my patch.
And I think about a patch again and send it.

Best Regards,
Hideo Yamauchi.

>
> Cheers,
>
> Dejan
>
> >
> > Best Regards,
> > Hideo Yamauchi.
> >
> > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> >
> > > Hi Hideo-san,
> > >
> > > On Fri, Apr 06, 2012 at 10:50:39AM +0900, renayama19661014 [at] ybb wrote:
> > > > Hi Dejan,
> > > >
> > > > I change validate-all and want to change it to always carry out validate-all.
> > > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > > >
> > > > How do you think?
> > >
> > > Isn't it that validate-all may be really necessary only in the
> > > start action? The repeating monitor is scheduled only after a
> > > successful start.
> > >
> > > Cheers,
> > >
> > > Dejan
> > >
> > > > Best Regards,
> > > > Hideo Yamauchi.
> > > >
> > > > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > >
> > > > > Hi Hideo-san,
> > > > >
> > > > > On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> > > > > > Hi Dejan,
> > > > > >
> > > > > > I agree to your patch.
> > > > >
> > > > > Thank you for the reply.
> > > > >
> > > > > BTW, the monitor was shamelessly stolen from Vladislav.
> > > > >
> > > > > Applied.
> > > > >
> > > > > ocft test passed (after some struggle and eventually fixing the
> > > > > ocft source).
> > > > >
> > > > > Cheers,
> > > > >
> > > > > Dejan
> > > > >
> > > > > > Best Regards,
> > > > > > Hideo Yamauchi.
> > > > > >
> > > > > > --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > > > >
> > > > > > > Hi all,
> > > > > > >
> > > > > > > This is a proposed set of two patches which would eliminate use
> > > > > > > of LVM commands in the monitor path. We already discussed the
> > > > > > > issue elsewhere and I don't see any point in keeping
> > > > > > > vgck/vgdisplay given that they don't result in better monitoring
> > > > > > > under normal circumstances. And if the circumstances are such
> > > > > > > that the new monitoring fails, I think that there'll be many
> > > > > > > more problems on the node than a failed volume group.
> > > > > > >
> > > > > > > Cheers,
> > > > > > >
> > > > > > > Dejan
> > > > > > >
> > > > > > _______________________________________________________
> > > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > > Home Page: http://linux-ha.org/
> > > > > _______________________________________________________
> > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > Home Page: http://linux-ha.org/
> > > > >
> > > > _______________________________________________________
> > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > Home Page: http://linux-ha.org/
> > >
> > _______________________________________________________
> > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > Home Page: http://linux-ha.org/
>
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/


dejan at suse

Apr 10, 2012, 12:53 AM

Post #10 of 11 (933 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Hideo-san,

On Tue, Apr 10, 2012 at 12:43:00PM +0900, renayama19661014 [at] ybb wrote:
> Hi Dejan,
>
> Thank you for comments.
>
>
> > Hi Hideo-san,
> >
> > On Mon, Apr 09, 2012 at 09:18:07AM +0900, renayama19661014 [at] ybb wrote:
> > > Hi Dejan,
> > >
> > > Thank you for comments.
> > >
> > > > > I change validate-all and want to change it to always carry out validate-all.
> > > > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > > > >
> > > > > How do you think?
> > > >
> > > > Isn't it that validate-all may be really necessary only in the
> > > > start action? The repeating monitor is scheduled only after a
> > > > successful start.
> > >
> > > It may be surely necessary as you say.
> > > However, I think validate-all to unify it so that it is always carried out.
> >
> > But why?
>
> There is the resource to carry out validate-all every time a lot.
> We wish it becomes LVM in the same way.

That's not a good reason. Testing if binaries exist on every
monitor operation really doesn't make much sense. Why would you
expect programs to start disappearing? And if they do, we may
have a much more serious problem to deal with.

Cheers,

Dejan

> > > How about what the check of vgck/vgdisplay chooses it in a parameter and can carry out?
> >
> > Again, why? It doesn't make any difference for a running
> > resource? We may do this before the start operation, of course.
>
> My correction is different from original LVM in big validate-all.
>
> There were many mistakes to my patch.
> And I think about a patch again and send it.
>
> Best Regards,
> Hideo Yamauchi.
>
> >
> > Cheers,
> >
> > Dejan
> >
> > >
> > > Best Regards,
> > > Hideo Yamauchi.
> > >
> > > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> > >
> > > > Hi Hideo-san,
> > > >
> > > > On Fri, Apr 06, 2012 at 10:50:39AM +0900, renayama19661014 [at] ybb wrote:
> > > > > Hi Dejan,
> > > > >
> > > > > I change validate-all and want to change it to always carry out validate-all.
> > > > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > > > >
> > > > > How do you think?
> > > >
> > > > Isn't it that validate-all may be really necessary only in the
> > > > start action? The repeating monitor is scheduled only after a
> > > > successful start.
> > > >
> > > > Cheers,
> > > >
> > > > Dejan
> > > >
> > > > > Best Regards,
> > > > > Hideo Yamauchi.
> > > > >
> > > > > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > > >
> > > > > > Hi Hideo-san,
> > > > > >
> > > > > > On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> > > > > > > Hi Dejan,
> > > > > > >
> > > > > > > I agree to your patch.
> > > > > >
> > > > > > Thank you for the reply.
> > > > > >
> > > > > > BTW, the monitor was shamelessly stolen from Vladislav.
> > > > > >
> > > > > > Applied.
> > > > > >
> > > > > > ocft test passed (after some struggle and eventually fixing the
> > > > > > ocft source).
> > > > > >
> > > > > > Cheers,
> > > > > >
> > > > > > Dejan
> > > > > >
> > > > > > > Best Regards,
> > > > > > > Hideo Yamauchi.
> > > > > > >
> > > > > > > --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > > > > >
> > > > > > > > Hi all,
> > > > > > > >
> > > > > > > > This is a proposed set of two patches which would eliminate use
> > > > > > > > of LVM commands in the monitor path. We already discussed the
> > > > > > > > issue elsewhere and I don't see any point in keeping
> > > > > > > > vgck/vgdisplay given that they don't result in better monitoring
> > > > > > > > under normal circumstances. And if the circumstances are such
> > > > > > > > that the new monitoring fails, I think that there'll be many
> > > > > > > > more problems on the node than a failed volume group.
> > > > > > > >
> > > > > > > > Cheers,
> > > > > > > >
> > > > > > > > Dejan
> > > > > > > >
> > > > > > > _______________________________________________________
> > > > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > > > Home Page: http://linux-ha.org/
> > > > > > _______________________________________________________
> > > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > > Home Page: http://linux-ha.org/
> > > > > >
> > > > > _______________________________________________________
> > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > Home Page: http://linux-ha.org/
> > > >
> > > _______________________________________________________
> > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > Home Page: http://linux-ha.org/
> >
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/


renayama19661014 at ybb

Apr 10, 2012, 9:49 PM

Post #11 of 11 (922 views)
Permalink
Re: LVM monitor change [In reply to]

Hi Dejan,

Thank you for comments.

> That's not a good reason. Testing if binaries exist on every
> monitor operation really doesn't make much sense. Why would you
> expect programs to start disappearing? And if they do, we may
> have a much more serious problem to deal with.

All right.

We withdraw this patch.
And let me discuss it when we review overall RA next again.

Many Thanks,
Hideo Yamauchi.


--- On Tue, 2012/4/10, Dejan Muhamedagic <dejan [at] suse> wrote:

> Hi Hideo-san,
>
> On Tue, Apr 10, 2012 at 12:43:00PM +0900, renayama19661014 [at] ybb wrote:
> > Hi Dejan,
> >
> > Thank you for comments.
> >
> >
> > > Hi Hideo-san,
> > >
> > > On Mon, Apr 09, 2012 at 09:18:07AM +0900, renayama19661014 [at] ybb wrote:
> > > > Hi Dejan,
> > > >
> > > > Thank you for comments.
> > > >
> > > > > > I change validate-all and want to change it to always carry out validate-all.
> > > > > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > > > > >
> > > > > > How do you think?
> > > > >
> > > > > Isn't it that validate-all may be really necessary only in the
> > > > > start action? The repeating monitor is scheduled only after a
> > > > > successful start.
> > > >
> > > > It may be surely necessary as you say.
> > > > However, I think validate-all to unify it so that it is always carried out.
> > >
> > > But why?
> >
> > There is the resource to carry out validate-all every time a lot.
> > We wish it becomes LVM in the same way.
>
> That's not a good reason. Testing if binaries exist on every
> monitor operation really doesn't make much sense. Why would you
> expect programs to start disappearing? And if they do, we may
> have a much more serious problem to deal with.
>
> Cheers,
>
> Dejan
>
> > > > How about what the check of vgck/vgdisplay chooses it in a parameter and can carry out?
> > >
> > > Again, why? It doesn't make any difference for a running
> > > resource? We may do this before the start operation, of course.
> >
> > My correction is different from original LVM in big validate-all.
> >
> > There were many mistakes to my patch.
> > And I think about a patch again and send it.
> >
> > Best Regards,
> > Hideo Yamauchi.
> >
> > >
> > > Cheers,
> > >
> > > Dejan
> > >
> > > >
> > > > Best Regards,
> > > > Hideo Yamauchi.
> > > >
> > > > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > >
> > > > > Hi Hideo-san,
> > > > >
> > > > > On Fri, Apr 06, 2012 at 10:50:39AM +0900, renayama19661014 [at] ybb wrote:
> > > > > > Hi Dejan,
> > > > > >
> > > > > > I change validate-all and want to change it to always carry out validate-all.
> > > > > > I abolish vgck/vgdisplay carried out in validate-all and intend to make only the check of the parameter simply.
> > > > > >
> > > > > > How do you think?
> > > > >
> > > > > Isn't it that validate-all may be really necessary only in the
> > > > > start action? The repeating monitor is scheduled only after a
> > > > > successful start.
> > > > >
> > > > > Cheers,
> > > > >
> > > > > Dejan
> > > > >
> > > > > > Best Regards,
> > > > > > Hideo Yamauchi.
> > > > > >
> > > > > > --- On Fri, 2012/4/6, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > > > >
> > > > > > > Hi Hideo-san,
> > > > > > >
> > > > > > > On Thu, Apr 05, 2012 at 11:32:05AM +0900, renayama19661014 [at] ybb wrote:
> > > > > > > > Hi Dejan,
> > > > > > > >
> > > > > > > > I agree to your patch.
> > > > > > >
> > > > > > > Thank you for the reply.
> > > > > > >
> > > > > > > BTW, the monitor was shamelessly stolen from Vladislav.
> > > > > > >
> > > > > > > Applied.
> > > > > > >
> > > > > > > ocft test passed (after some struggle and eventually fixing the
> > > > > > > ocft source).
> > > > > > >
> > > > > > > Cheers,
> > > > > > >
> > > > > > > Dejan
> > > > > > >
> > > > > > > > Best Regards,
> > > > > > > > Hideo Yamauchi.
> > > > > > > >
> > > > > > > > --- On Thu, 2012/4/5, Dejan Muhamedagic <dejan [at] suse> wrote:
> > > > > > > >
> > > > > > > > > Hi all,
> > > > > > > > >
> > > > > > > > > This is a proposed set of two patches which would eliminate use
> > > > > > > > > of LVM commands in the monitor path. We already discussed the
> > > > > > > > > issue elsewhere and I don't see any point in keeping
> > > > > > > > > vgck/vgdisplay given that they don't result in better monitoring
> > > > > > > > > under normal circumstances. And if the circumstances are such
> > > > > > > > > that the new monitoring fails, I think that there'll be many
> > > > > > > > > more problems on the node than a failed volume group.
> > > > > > > > >
> > > > > > > > > Cheers,
> > > > > > > > >
> > > > > > > > > Dejan
> > > > > > > > >
> > > > > > > > _______________________________________________________
> > > > > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > > > > Home Page: http://linux-ha.org/
> > > > > > > _______________________________________________________
> > > > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > > > Home Page: http://linux-ha.org/
> > > > > > >
> > > > > > _______________________________________________________
> > > > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > > > Home Page: http://linux-ha.org/
> > > > >
> > > > _______________________________________________________
> > > > Linux-HA-Dev: Linux-HA-Dev [at] lists
> > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > > > Home Page: http://linux-ha.org/
> > >
> _______________________________________________________
> Linux-HA-Dev: Linux-HA-Dev [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
>
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

Linux-HA dev RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.