how does xtrabackup copy data files?

1 view
Skip to first unread message

Baron Schwartz

unread,
Nov 8, 2010, 5:33:16 PM11/8/10
to perco...@googlegroups.com
A few questions that I had a hard time figuring out from the source:

- are the files copied in page-sized blocks?
- are the page-sized blocks checksummed during the copy to catch
corruption?
- is there a retry in case there is corruption, such as reading a page
at the same time it's being written? Does that matter?
- are there any other internal details that are interesting or
important to know?

Peter Zaitsev

unread,
Nov 8, 2010, 8:28:57 PM11/8/10
to perco...@googlegroups.com
Baron,

As far as  I remember looking at Strace it uses 1MB  reads...   But yes details of this
should be documented. 


--
You received this message because you are subscribed to the Google Groups "Percona Dev" group.
To post to this group, send email to perco...@googlegroups.com.
To unsubscribe from this group, send email to percona-dev...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/percona-dev?hl=en.




--
Peter Zaitsev, CEO, Percona Inc.
Tel: +1 888 401 3401 ext 501   Skype:  peter_zaitsev
24/7 Emergency Line +1 888 401 3401 ext 911

Percona Training Workshops
http://www.percona.com/training/

Baron Schwartz

unread,
Nov 8, 2010, 9:14:25 PM11/8/10
to perco...@googlegroups.com
Hi Peter,

I'm not sure why I didn't think to use strace. You're right, it's 1MB
at a time. And then I also found the right spot in the source code,
which somehow I missed before. It iterates over the 1MB buffer a page at
a time, and checks for page corruption, and retries up to 10 times for
each page. It skips this check on the doublewrite buffer.

On Mon, Nov 8 at 17:28 Peter Zaitsev wrote:
> Baron,
>
> As far as I remember looking at Strace it uses 1MB reads... But yes
> details of this
> should be documented.
>
> On Mon, Nov 8, 2010 at 2:33 PM, Baron Schwartz <ba...@percona.com> wrote:
>
> > A few questions that I had a hard time figuring out from the source:
> >
> > - are the files copied in page-sized blocks?
> > - are the page-sized blocks checksummed during the copy to catch
> > corruption?
> > - is there a retry in case there is corruption, such as reading a page
> > at the same time it's being written? Does that matter?
> > - are there any other internal details that are interesting or
> > important to know?
> >
> > --
> > You received this message because you are subscribed to the Google Groups
> > "Percona Dev" group.
> > To post to this group, send email to perco...@googlegroups.com.
> > To unsubscribe from this group, send email to

> > percona-dev...@googlegroups.com<percona-dev%2Bunsu...@googlegroups.com>


> > .
> > For more options, visit this group at
> > http://groups.google.com/group/percona-dev?hl=en.
> >
> >
>
>


--
Chief Performance Architect at Percona <http://www.percona.com/>
+1 (888) 401-3401 x507
Calendar: <http://bit.ly/baron-percona-cal> (Eastern Time)
Flat-rate 24x7 support for MySQL <http://tinyurl.com/2fcmn34>

Reply all
Reply to author
Forward
0 new messages