Re: [ANNOUNCE][PATCH 2.6] md: persistent (file-backed) bitmap and async writes

[Paul Clements <paul.clements@steeleye.com>]

Neil Brown wrote:

> The changes to mdadm need a bit of work.
>  
> You have added "--persistent" and "--non-persistent" flags for
> --create.  This is wrong.
> --create always uses persistent superblocks.
> --build makes arrays without persistent superblocks.

Yes, I think --build didn't use to work for raid1, though, which is why 
I went the create route. I'll fix this.

> I don't think I like --create-bitmap.  A bitmap file should always be
> created in the context of a particular array (partly so that the size
> and uuid can be set correctly).  I think I would like to bitmap to be
> specified as a "--bitmap=filename" option to --create, --build, or
> --grow.  I haven't thought this through in great detail yet, but that
> is my leaning.

That all sounds fine except for --build. How do we know if we're doing 
the initial build or just a rebuild. We need to know this so we know 
whether to initialize the bitmap or simply use it. So I think we still 
need --create-bitmap for the build command (and then build will just use 
the bitmap, not initialize it). For --create I can make it so that the 
bitmap is always initialized using the array parameters, so no separate 
--create-bitmap will be needed. I'll need to work out the conflict that 
this will create with the --chunk option, though, since --chunk is the 
array chunk size for --create and --chunk is the bitmap chunk size for 
--bitmap-create. Any suggestions?

> --examine-bitmap is a bit of a pain too. 

-X is usually what I use...that's a lot easier to remember and type...

> I think I would like
> --examine to figure out what it has been given to look at, and report
> on whatever it finds.

Well, that gets into a little bit of black magic when you're talking 
about examining a disk or partition, which could conceivably contain 
both an md superblock at the end and a bitmap at the front. How do we 
know which one it is "supposed" to be? Or do you just want examine to 
print information for both? That would seem a little confusing to me 
(granted this would only happen when a disk changed roles between bitmap 
and array component).

--
Paul