Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: DRBD: Users

"dd" meta-data when DRBD's role is Primary or Secondary

 

 

DRBD users RSS feed   Index | Next | Previous | View Threaded


tsukishima.ha at gmail

Jul 27, 2010, 7:06 AM

Post #1 of 2 (341 views)
Permalink
"dd" meta-data when DRBD's role is Primary or Secondary

Hi,

I'm trying to confirm whether DRBD + Heartbeat cluster can put up with
the following scenario.

(1) startup DRBD + Heartbeat
Active node's role: Primary
Standby node's role: Secondary

(2) crash the meta-data
call "dd" command on the both of Active and Standby node like this;
They have internal meta-data on /dev/cciss/c0d0p6.

# dd if=/dev/zero of=/dev/cciss/c0d0p6 count=100000

after that, nothing has happened.
Active node kept running its service,
and after stopping Active node, Standby node took over the service.
This behavior is not a problem,
but I just want to know whether DRBD can keep its meta-data on
memory(?) if DRBD's role is Primary or Secondary.

If I down the DRBD's resource and "dd" the meta-data,
it seemed that meta-data was crashed, and I needed to recreate it.
# drbdadm down all
# dd if=/dev/zero of=/dev/cciss/c0d0p6 count=100000
# drbdadm up all
---> Failed!
# drbdadm create-md all

Thanks,
Junko

NTT DATA INTELLILINK CORPORATION
_______________________________________________
drbd-user mailing list
drbd-user [at] lists
http://lists.linbit.com/mailman/listinfo/drbd-user


lars.ellenberg at linbit

Jul 27, 2010, 7:38 AM

Post #2 of 2 (330 views)
Permalink
Re: "dd" meta-data when DRBD's role is Primary or Secondary [In reply to]

On Tue, Jul 27, 2010 at 11:06:42PM +0900, Junko IKEDA wrote:
> Hi,
>
> I'm trying to confirm whether DRBD + Heartbeat cluster can put up with
> the following scenario.
>
> (1) startup DRBD + Heartbeat
> Active node's role: Primary
> Standby node's role: Secondary
>
> (2) crash the meta-data
> call "dd" command on the both of Active and Standby node like this;
> They have internal meta-data on /dev/cciss/c0d0p6.
>
> # dd if=/dev/zero of=/dev/cciss/c0d0p6 count=100000
>
> after that, nothing has happened.

assuming you have a partition table on sda, and you do
grep sda /proc/partitions
dd if=/dev/zero of=/dev/sda bs=4k count=1
grep sda /proc/partitions
"nothing has happened" either,
but your next reboot won't get very far.

Assuming you have some decent ram, and run
find /mntpoint -ls > /dev/null
to get a not too large file system dentry and inodes into the cache.
then you dd if=/dev/zero of=/underlying-block-device,
you will still be able to cd, and maybe even ls -l and stuff,
as long as it stays in the cache.

> Active node kept running its service,
> and after stopping Active node, Standby node took over the service.
> This behavior is not a problem,
> but I just want to know whether DRBD can keep its meta-data on
> memory(?) if DRBD's role is Primary or Secondary.

No. It knows about its _state_, and on state changes,
writes (part of) that state to stable storage.

> If I down the DRBD's resource and "dd" the meta-data,
> it seemed that meta-data was crashed, and I needed to recreate it.
> # drbdadm down all
> # dd if=/dev/zero of=/dev/cciss/c0d0p6 count=100000
> # drbdadm up all
> ---> Failed!
> # drbdadm create-md all

So what? What are you going to prove here? That you are able to use dd
to damage your system? No news in that.

Sorry, but this is ridiculous.
I really don't see where this should be going.

--
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.
__
please don't Cc me, but send to list -- I'm subscribed
_______________________________________________
drbd-user mailing list
drbd-user [at] lists
http://lists.linbit.com/mailman/listinfo/drbd-user

DRBD users RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.