open-sharedroot-devel

[OSR-devel] Two node cluster - possible bug in /usr/bin/cl_checknodes

[<go...@bo...>]

I've been trying to track down the issue of a cluster with:
two_node="1"
expected_vodes="1"
not coming up as quorate with just one node, and as far as I can tell, it 
is because /usr/bin/cl_checknodes is returning the wrong value.

I think this bit around line 72 is causing the problem:

if len(nodeelements) == 1:
     quorum=1
else:
     quorum=len(nodeelements)/2+1

Should that not instead be something like:

if len(nodeelements) <= 2:
     quorum=1
else:
     quorum=len(nodeelements)/2+1


Please advise.

Gordan

Re: [OSR-devel] Two node cluster - possible bug in /usr/bin/cl_checknodes

[Marc G. <gr...@at...>]

On Monday 11 February 2008 17:33:15 go...@bo... wrote:
> I've been trying to track down the issue of a cluster with:
> two_node="1"
> expected_vodes="1"
> not coming up as quorate with just one node, and as far as I can tell, it
> is because /usr/bin/cl_checknodes is returning the wrong value.
>
> I think this bit around line 72 is causing the problem:
>
> if len(nodeelements) == 1:
>      quorum=1
> else:
>      quorum=len(nodeelements)/2+1
>
> Should that not instead be something like:
>
> if len(nodeelements) <= 2:
>      quorum=1
> else:
>      quorum=len(nodeelements)/2+1
>
>
> Please advise.
Basically that wasn't the idea behind it. We wanted to prevent a cluster with 
more then one node to not coming up with splitbrain when both nodes are 
powered up initially. It's basically a way to wait until both nodes are up 
and running. We didn't want to risk doublemounts for users not being very 
sensitive with booting machines or just first level supporters don't 
bothering much about data consistency. 

So for you I would be just add the bootoption quorumack if you don't have 
anybody else being able to reboot/fence your clusternodes without much 
experience.

Marc.
>
> Gordan
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Open-sharedroot-devel mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-sharedroot-devel



-- 
Gruss / Regards,

Marc Grimme
http://www.atix.de/               http://www.open-sharedroot.org/

Re: [OSR-devel] Two node cluster - possible bug in /usr/bin/cl_checknodes

[<go...@bo...>]

On Tue, 12 Feb 2008, Marc Grimme wrote:

>> I've been trying to track down the issue of a cluster with:
>> two_node="1"
>> expected_vodes="1"
>> not coming up as quorate with just one node, and as far as I can tell, it
>> is because /usr/bin/cl_checknodes is returning the wrong value.
>>
>> I think this bit around line 72 is causing the problem:
>>
>> if len(nodeelements) == 1:
>>      quorum=1
>> else:
>>      quorum=len(nodeelements)/2+1
>>
>> Should that not instead be something like:
>>
>> if len(nodeelements) <= 2:
>>      quorum=1
>> else:
>>      quorum=len(nodeelements)/2+1
>>
>>
>> Please advise.
>
> Basically that wasn't the idea behind it. We wanted to prevent a cluster with
> more then one node to not coming up with splitbrain when both nodes are
> powered up initially. It's basically a way to wait until both nodes are up
> and running. We didn't want to risk doublemounts for users not being very
> sensitive with booting machines or just first level supporters don't
> bothering much about data consistency.
>
> So for you I would be just add the bootoption quorumack if you don't have
> anybody else being able to reboot/fence your clusternodes without much
> experience.

OK, that makes sense. Where should the "quorumack" option be? In 
cluster.conf? If so, which tag/section? Or is it a kernel boot parameter 
option?

Gordan

Re: [OSR-devel] Two node cluster - possible bug in /usr/bin/cl_checknodes

[Marc G. <gr...@at...>]

On Tuesday 12 February 2008 12:07:11 go...@bo... wrote:
> On Tue, 12 Feb 2008, Marc Grimme wrote:
> >> I've been trying to track down the issue of a cluster with:
> >> two_node="1"
> >> expected_vodes="1"
> >> not coming up as quorate with just one node, and as far as I can tell,
> >> it is because /usr/bin/cl_checknodes is returning the wrong value.
> >>
> >> I think this bit around line 72 is causing the problem:
> >>
> >> if len(nodeelements) == 1:
> >>      quorum=1
> >> else:
> >>      quorum=len(nodeelements)/2+1
> >>
> >> Should that not instead be something like:
> >>
> >> if len(nodeelements) <= 2:
> >>      quorum=1
> >> else:
> >>      quorum=len(nodeelements)/2+1
> >>
> >>
> >> Please advise.
> >
> > Basically that wasn't the idea behind it. We wanted to prevent a cluster
> > with more then one node to not coming up with splitbrain when both nodes
> > are powered up initially. It's basically a way to wait until both nodes
> > are up and running. We didn't want to risk doublemounts for users not
> > being very sensitive with booting machines or just first level supporters
> > don't bothering much about data consistency.
> >
> > So for you I would be just add the bootoption quorumack if you don't have
> > anybody else being able to reboot/fence your clusternodes without much
> > experience.
>
> OK, that makes sense. Where should the "quorumack" option be? In
> cluster.conf? If so, which tag/section? Or is it a kernel boot parameter
> option?
It's only a bootparm ;-).
>
> Gordan
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Open-sharedroot-devel mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-sharedroot-devel



-- 
Gruss / Regards,

Marc Grimme
Phone: +49-89 452 3538-14
http://www.atix.de/               http://www.open-sharedroot.org/

**
ATIX Informationstechnologie und Consulting AG
Einsteinstr. 10 
85716 Unterschleissheim
Deutschland/Germany

Phone: +49-89 452 3538-0
Fax:   +49-89 990 1766-0

Registergericht: Amtsgericht Muenchen
Registernummer: HRB 168930
USt.-Id.: DE209485962

Vorstand: 
Marc Grimme, Mark Hlawatschek, Thomas Merz (Vors.)

Vorsitzender des Aufsichtsrats:
Dr. Martin Buss

Re: [OSR-devel] Two node cluster - possible bug in /usr/bin/cl_checknodes

[<go...@bo...>]

On Tue, 12 Feb 2008, Marc Grimme wrote:

>>>> I've been trying to track down the issue of a cluster with:
>>>> two_node="1"
>>>> expected_vodes="1"
>>>> not coming up as quorate with just one node, and as far as I can tell,
>>>> it is because /usr/bin/cl_checknodes is returning the wrong value.
>>>>
>>>> I think this bit around line 72 is causing the problem:
>>>>
>>>> if len(nodeelements) == 1:
>>>>      quorum=1
>>>> else:
>>>>      quorum=len(nodeelements)/2+1
>>>>
>>>> Should that not instead be something like:
>>>>
>>>> if len(nodeelements) <= 2:
>>>>      quorum=1
>>>> else:
>>>>      quorum=len(nodeelements)/2+1
>>>>
>>>>
>>>> Please advise.
>>>
>>> Basically that wasn't the idea behind it. We wanted to prevent a cluster
>>> with more then one node to not coming up with splitbrain when both nodes
>>> are powered up initially. It's basically a way to wait until both nodes
>>> are up and running. We didn't want to risk doublemounts for users not
>>> being very sensitive with booting machines or just first level supporters
>>> don't bothering much about data consistency.
>>>
>>> So for you I would be just add the bootoption quorumack if you don't have
>>> anybody else being able to reboot/fence your clusternodes without much
>>> experience.
>>
>> OK, that makes sense. Where should the "quorumack" option be? In
>> cluster.conf? If so, which tag/section? Or is it a kernel boot parameter
>> option?
>
> It's only a bootparm ;-).

So, just something like this in grub.conf is sufficient?
kernel /2.6.18-53.1.6.el5/vmlinuz ro quorumack

Gordan