say after all transactions in the log are replayed, the log space may be
reclaimed. if the log space reclaimation writes hit the persistent
storage earlier than the transaction replay writes, and a crash happens in
between, then we lose commited transactions in the log. so presumably
there should be some sync/barrier points in log replaying. my question
is, what's the syncing mechanism in xfs log replay
-junfeng
On Fri, 27 Aug 2004, Eric Sandeen wrote:
> I don't think we -do- guarantee ordered writes during journal replay...
> any context for this question?
>
> All of the original design docs are at
> http://oss.sgi.com/projects/xfs/design_docs/xfsdocs93_pdf/
>
> there are also talks & papers at
> http://oss.sgi.com/projects/xfs/publications.html which may have some
> useful info.
>
> -Eric
>
> Junfeng Yang wrote:
> > hi Eric,
> >
> > a quick question for xfs journal replay: what sync functions do you use
> > to guanrentee ordered writes when replaying the journal at mount time?
> > for ext3 they use wait_on_buffer. i didn't see any wait_on_buffer calls
> > in xfs code. is it ic_writesema?
> >
> > also, it'll be great if you can point me to any xfs logging documents.
> >
> > thanks,
> > -Junfeng
>
|