[Box Backup-dev] Re: [Box Backup] How to move boxbackup store to a larger disk

[Ben Summers]

Chris Wilson wrote:
>
> Hi Ben,
>
> On Sun, 14 Sep 2008, Ben Summers wrote:
>
>> The container ID is set on the client side when the stored file is  
>> originally
>> created.
>
> Is it part of the file's header on the server then?

Yes.

See BackupStoreFileEncodeStream::Setup() and file_StreamFormat.

>
>
>> It's not used for anything except store reconstruction, so it seemed
>> a bit of a waste of time to update it when files were moved.
>
> I guess if we had to use it for reconstruction, and it was wrong, then
> we'd end up putting the file back into the wrong directory, or not  
> knowing
> where it should be at all?

Yes, that's correct. But after the first sync from the backup client,  
it would be moved to the correct position. Failure of the server and  
backed up server should be unlikely to occur at the same time.

>
>
>> If it were to be changed, the server would have to do it, which given
>> the restrictive RaidFile interface would require rewriting the file  
>> in
>> it's entirety.
>
> Can we fix the RaidFile interface? Especially as actual 3-stripe  
> RAID is
> going away, we could make this a no-op on 3-stripe RAID and trivial on
> non-RAID?

I wouldn't say "fix" is quite the word, you'd need to change the way  
it worked underneath. I suspect it would be better to remove RaidFile  
complete, and do some sort of automated change to FileStreams.

>
>
>> Maybe a reasonable solution is to simply remove that "warning"?
>
> If it's too hard to actually repair the problem then I fully support  
> doing
> that to ease needless user fears.

Yes. Or at a very verbose logging method and a rewritten message. It's  
more useful for developers than anyone else.

Ben