[Linux-cluster] Re: Linux-cluster Digest, Preventing LVM from concurrent access

[Rafael Micó Miranda]

Hi Edson

El lun, 24-08-2009 a las 12:13 -0300, Edson Marquezani Filho escribió:

[...]
> 
> Hum... that's good too. By the way, I was concerned about that
> requisite, which is absolutely not much pleaseant to do.
> 
> Anyway, I'm trying to set up your script here, but I realized that
> actualy I still don't know how to do that. =/
> (I'm using that tool called Luci, which looks good, specialy to
> someone that knows almost nothing about cluster. =)


I know about Luci and Ricci. I'm pretty sure you would not be able to
use my script with then, either with the system-config-cluster tool
which i have tested, because the "don't know" about this resource
script. You will need to use it, meanwhile this script gets into the
project or not, configuring the cluster.conf file by hand.

> Ok, I have defined a script resource with the path of your script, and
> I guess that now I have to create a service, that's it?
> But how, if I don't have a single service to be watched. (Actualy, my
> service is my whole server, we could say.)


You dont need to create a new script that uses mine. Take a look to the
cluster.conf file and the RedHat manuals about Cluster Administration
and configure my resource script as part of a service, as your are
suspecting.

You need to define a service the cluster has to handle. You plan to run
Xen Virtual Machines, don't you? You'll need a service, at least, formed
by:

1.- An IP Address where the service is offered
2.- A shared storage where the virtual machines lay
3.- A "binary" or "system call out" to start the Xen service when the
two preceding items are taken. 

> Could you give me a configuration example? I'm really confused here yet.
> 
> Wait, should I write a script for this task and use it on service
> configuration? I have tought about a script that always returns OK to
> verify-all, and status, and lauch VMs for start. Does it make any
> sense?
> 

I provide you as an attached file an example cluster.conf file I have
just prepared to show you as an example of lvm-cluster usage. It is for
a NFS export service to a client. Fencing is not configured, but qdisk
is.

> > 2.- Communications interruption and split-brain: you should use some
> > fencing system to make the 2nd node "kill" the main one and take the
> > resources.
> 
> Ok, I have understood that there's no way to avoid that problem unless
> turning off a server. Anyway, this is acceptable, once I don't expect
> such a situation happening oftenly.
> 

Yes, you should always use fencing when there is an attached shared
storage.

> > Having the main node booted back again AND the comms not reestablished
> > will cause you a split-brain again, with a new fencing action. I usually
> > do not set cluster services (CMAN, RGMANAGER, etc.) to be automatically
> > started on the nodes. I mean they must be re-run and activated by an
> > admin.
> >
> > In the other hand you can use a Quorum Disk, provided my CMAN too, as
> > the 3rd vote on your cluster and keep the quorum. I use both approaches
> > at the same time (no auto-start of cluster services and qdisk).
> >
> > 3.- Communications channels: there's no way to set on CMAN more than one
> > comm channel, but you can set more than one network interface to be part
> > of that channel, via bonding. I think this is one of the lacks of CMAN
> > actually.
> >
> > I hope this helps. Just ask whatever you want.
> >
> 
> Thank you, very much.
> 
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cluster

Thanks for your interest. Cheers,

Rafael

-- 
Rafael Micó Miranda
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cluster.conf.example
Type: application/xml
Size: 1661 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20090824/41b33968/attachment.wsdl>