Re: [BUG] non-metadata arrays cannot use more than 27 component devices

[Anthony Youngman <antlists@youngman.org.uk>]

On 01/03/17 18:13, Phil Turmel wrote:
> On 03/01/2017 12:23 PM, Phil Turmel wrote:
>> I strongly disagree.  This procedure, as shown, is an admin cock-up:
>>
>>> mdadm --build /dev/mdbackup --device-count 2 /dev/md/home missing
>>> ... hotplug sd-big ...
>>> madam /dev/mdbackup --add /dev/sd-big
>>> ... wait for sync to finish ...
>>> mdadm --stop mdbackup
>>> ... unplug sd-big ...
>
> One more point.  The above is functionally identical in every respect
> to just:
>
> # dd if=/dev/md/home of=/dev/sd-big bs=1M
>
> Why are you bothering to --build an array?
>
Because - and this is a point the kernel guys seem to forget - the whole 
point of having a computer system is TO RUN APPLICATIONS, not to run an OS.

As it is, you picked up on the fatal flaw I'd spotted, namely that if 
"home" is mounted, "backup" is going to be corrupt :-( Defeating the 
entire purpose of my idea, which was to back up a running system without 
the need to take down the system to ensure integrity.

I work with a database that, not unreasonably, seeks to cache loads of 
stuff in RAM. I've come across far too many horror stories of corrupt 
backups because the database hadn't flushed its buffers to the OS, so 
all the database files on disk were inconsistent, giving a corrupt 
backup. So the idea was set up the mirror, flush/quiesce the database, 
break the mirror, wake up the database. System disabled for a matter of 
seconds.

It's all very well saying lvm was created with this in mind, but if the 
system wasn't installed with this originally in mind, you're up a gum 
tree. My home system is raid but not lvm, for example - how do I back up 
the system while it's live? (In reality, I don't care :-)

IF it didn't have that fatal flaw, my idea would have been able to back 
up any system. Oh well, it's flawed, time to drop it :-(

Cheers,
Wol