Subject: Re: [vserver] chroot barrier problem when consolidating var and etc
From: Martin Fick <mogulguy@yahoo.com>
Date: Fri, 3 Oct 2008 10:49:50 -0700 (PDT)

--- On Fri, 10/3/08, randall <randall@songshu.org> wrote:

> it took me some time to get it but i do have a working set
> up right now, 
> i could not find many complete examples on how to do this
> with Vserver 
> so i thought my notes might be a nice addition to the wiki.

This is very nice, I think that it will be very helpful.  
My failover page was abstract and lacked much in the
way of examples.  On the other hand, your page 
provides some good concrete examples.  I think that
both pages complement each other well.  I would suggest 
linking to your page from the failover page.

 
> http://linux-vserver.org/Getting_high_with_lenny#start.2Fstop_script

2 suggestions:  

A) I would suggest not putting the ARP hack 
directly in the OCF script.  The reason I say 
that is because if you ever want to start 
and stop the vserver manually this piece will be
missing.  Instead, I would suggest moving the one
liner to a script of its own and make it a part
of the vserver startup process itself.  This way
it will get run whenever the vserver is started
whether it be manually or with heartbeat.

B) Perhaps instead of hacking the name into the
OCF script and duplicating it, it might be better
to place the unaltered (unless there are even 
better improvements to the script) OCF agent
in its normal R2 location, and to simply create
an additional custom script that calls it for 
each of the R1 resources that you need.  That
custom script could set the instance name 
variable and then call the OCF agent.  This
would mean less deviation, less duplication
and make the customization more explicit.

Later,

-Martin