[Ocfs2-users] 2 node cluster reboot

[Sunil Mushran]

For starters, (RH)EL5 2.6.18 is not the same as 2.6.18 mainline.

Secondly, we backport the latest file system to enterprise
kernels, (rh)el and sles. We do not backport to older mainline
kernels. The fs in mainline 2.6.18 is very very old.

If you are on an enterprise kernel, you could go with ocfs2 1.2.9
or ocfs2 1.4.1. ocfs2 1.4.1 is newer and thus has more features.
If you are on (RH)EL5 U2+, go with 1.4.1.

Debian uses the mainline kernel. Meaning you don't get a choice
of the fs version. Use the one the distro ships.

For more, refer to the ocfs2 1.4 user's guide.

Dante Garro wrote:
> Sunil, I'm confused.
>
> You advice me to get e newer kernel version, 
> xen doesn't work beyond 2.6.18 kernel version, so I'm planning to leave it
> and migrate to a regular 
> environment in this way I could obtain up to 2.6.26 kernel version from
> Debian Backports, but when I went to ocfs site to look for a suitable ocfs2
> version for that kernel, I can't find any package beyond 2.6.18.92!
>
> Following my questions:
>
> - why seems old to you my kernel version?
> - what could be more stable? 
> 	1 Centos 5.2 (kernel 2.6.18.92.1.13) with ocfs2 1.2 or with ocfs2
> 1.4?
> 	2 Debian's 2.6.26 kernel with ocfs2 1.2 or with ocfs2 1.4?
> - if I switch to Debian's 2.6.26, what version of ocfs2 can I use?
>
> Thank you, 
>
> Dante
>
>
>
>
> -----Mensaje original-----
> De: Dante Garro 
> Enviado el: martes, 23 de septiembre de 2008 21:38
> Para: 'Sunil Mushran'
> CC: 'ocfs2-users at oss.oracle.com'
> Asunto: RE: [Ocfs2-users] 2 node cluster reboot
>
> Thanks for your promptly answer.
> I know my kernel it's a bit old, but I'm not feel so brave to try a xen's
> kernel compilation....
>
> I've talked about the case of a node crash, it never will shutdown the
> service orderly. In this case a with just one node running are there a way
> to keep it running avoiding panicing itself?
>
> Dante
>
>
>
> -----Mensaje original-----
> De: Sunil Mushran [mailto:sunil.mushran at oracle.com] Enviado el: Martes, 23
> de Septiembre de 2008 01:52 p.m.
> Para: Dante Garro
> CC: 'ocfs2-users at oss.oracle.com'
> Asunto: Re: [Ocfs2-users] 2 node cluster reboot
>
> Yes, that is the whole point of a clustered file system.
>
> But for it to work correctly, it does require proper configuration.
> For example, correct service shutdown ordering. The FAQ explains one such
> issue that you are probably encountering.
>
> BTW, this version is 2.5yrs old. Upgrade to kernel 2.6.21+. Atleast.
>
> Dante Garro wrote:
>   
>> Thanks Sunil
>>
>> I'm running Debian 4 (2.6.18-6-xen-686) with ocfs2 version 1.2.1-1.3
>>
>> I understanded your point.
>> But in my case (2 node cluster) I hope if one node crashed the other 
>> keeps servicing the cluster services until I have it repaired/serviced .
>> Are there a solution for my case with ocfs?
>>
>> Dante
>>
>>
>>
>> -----Mensaje original-----
>> De: Sunil Mushran [mailto:sunil.mushran at oracle.com] Enviado el: 
>> Martes, 02 de Septiembre de 2008 05:09 p.m.
>> Para: Dante Garro
>> CC: 'ocfs2-users at oss.oracle.com'
>> Asunto: Re: [Ocfs2-users] 2 node cluster reboot
>>
>> Check out qs 80/81 in the faq.
>> http://oss.oracle.com/projects/ocfs2/dist/documentation/v1.2/ocfs2_faq
>> .html#
>> QUORUM
>>
>> which version, distro, etc.?
>>
>> Dante Garro wrote:
>>   
>>     
>>> Hi Sirs, I have a 2 nodes cluster.
>>>
>>> Seems (but I'm not sure)when I've rebooted one of nodes the other 
>>> reboots itself too.
>>>
>>> Is possible that ocfs services reboots the server that should stay
>>>       
> online?
>   
>>> How can I resolve it?
>>>
>>> Regards
>>>
>>> Dante
>>>
>>> _______________________________________________
>>> Ocfs2-users mailing list
>>> Ocfs2-users at oss.oracle.com
>>> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>>>   
>>>     
>>>       
>> _______________________________________________
>> Ocfs2-users mailing list
>> Ocfs2-users at oss.oracle.com
>> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>>   
>>