[Linux-ha-dev] Suggestion on disk based HBcomm plugin ?

[Lars Marowsky-Bree]

On 2007-06-05T12:08:17, Xinwei Hu <hxinwei at gmail.com> wrote:

>  Here's the updated version. It can work on a 2-vmware-nodes hb2 cluster
> with
> several dummy resources running for hours now. :)

Cool! Good work.

> Why do you need SHA inside the plugin!? The packets you receive are
> >already authenticated and signed.
> 
> I use multi-blocks for each message. In case of  half-written blocks, I use
> sha to compute checksum.

If you're using circular buffers with atomically updated pointers (after
the message write has completed, of course), you'll either see a full
message or none, so you don't need to verify the message integrity in
addition to that.

Even _if_ you see a half-written message or block, it doesn't matter.
The higher level authentication layers in heartbeat will simply discard
the message - it's just like any other form of corruption which
heartbeat can already deal with. So, I argue that this SHA code seems
unnecessary.

Thanks for getting started on this! There also was a discussion with
Harald Milz on the linux-ha list a few months (or maybe 1-2 years? ;-)
ago on this topic, which might bring up some points I have forgotten so
far.


Regards,
    Lars

-- 
Teamlead Kernel, SuSE Labs, Research and Development
SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde