Dear Tim,

Thanks for the info and valuable suggestions.

On Wed, May 20, 2009 at 10:18 AM, Tim Post <> wrote:

On Wed, 2009-05-20 at 09:40 +0530, er krishna wrote:

>         Stop vblade/vbladed on the storage server.
> How ? Currently I am doing ps-eaf | grep vblade ; & then I kill all
> the process; is there any other way ?

You could try the command pidof vblade. This will print all vblade PID's
to stdout.

so kill `pidof vblade` may be easier for you, however it is not
discriminate. In other words, you may want to stop just one vblade, not
all of them. Currently there is no easy way to keep track of which PID
is exporting what.

I have some patches that I'll send later this week that 'properizes'
vblade for use as a daemon. It then becomes easier to manage. The
changes are kind of invasive, so I'm guessing they can just live in

>>>> What about a new dameon, if we create vbladectl which will keep track of all the exported device based on  interface & path ( /dev/loop0 or eth0 ) release the attached resource or add new resource ? I think we can have a look for this implemenatation.

Ed, does it seems valuable or funny. A patch will be welcome if I try for this implementation. Just suggest.


> ok, so aoe-flush is necessary; Currently I am doing rmmod aoe ;
> modprobe aoe ; aoe-stat  & results are as per exptation. But aoe-flush
> seems better way.


> 1) On server ps -eaf | grep vblade;  find all the processes & kill
> those by kill -9.

Don't use -9, its not needed. Let vblade / vbladed handle the signals as
they are designed to do (i.e. clean up tasks, etc).

> 2) Run ssh command & login on each machine ( w/o passwd ) ; then
> execute aoe-flush command on each node on the network.

Yes, that is the simplest way. Puppet as others have mentioned would
also do the job.