rsync read block size

[Ming Zhang]

it become a BDD question now.. ;)

On Fri, May 22, 2009 at 5:37 PM, Matthias Schniedermeyer <ms at citd.de> wrote:
> On 22.05.2009 16:25, Ming Zhang wrote:
>> Hi All
>>
>> We want to use rsync to backup a live Berkley db to a remote site. BDB
>> has a requirement that read has to be in the unit of db page size. So
>> wonder how could we make sure that rsync can follow that? If we need
>> to change the code, where we should begin to look at? Thanks!
>
> The read block size is irrelevant, the problem you have is consistency.
>
> If you want a working copy of a database (and this goes for any
> database) you have to quiescent the database first. Otherwise the
> file(s) in the filesystem may not be in a consistent state.

if i stop the DB operation, then yes, block size does not matter any more.

but actually base on BDB document, you CAN copy the files, just need
to be careful on the order of files that get copied. i will use
db_hotbackup to figure out the file order and feed that to rsync as
files-from. now the block size question become relevant.




>
> Also you must either quiescent the database the WHOLE TIME you back it
> up, or you must use some form of snapshoting for the filesystem and then
> backup the snapshot. (quiescent -> Snapshot -> unquiescent)
> (Or use some other database depend backup procedure)
>
> How you quiescent a database is about as database dependend as it can
> get and personally i don't know how the procedure is for a BDB.
> But i guess(!) you need cooperation from the program(s) working with the
> db.
>
>
>
> Googling: backup berkeley db
> Appear to give answers, but i only took a quick glance.
>
>
>
> Bis denn
>
> --
> Real Programmers consider "what you see is what you get" to be just as
> bad a concept in Text Editors as it is in women. No, the Real Programmer
> wants a "you asked for it, you got it" text editor -- complicated,
> cryptic, powerful, unforgiving, dangerous.
>
>