[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Fatal crash on xen4.2 HVM + qemu-xen dm + NFS


--On 6 March 2013 20:01:53 -0500 Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx> 

Results are positive.


Does that mean some sort of switch to turn off O_DIRECT (or turning it off
by default) can go into qemu?

Thought it is interesting that the application is doing write and sync
(WS) for other blocks (like 2013633), which I thought would have
been the case if the file was opened with O_SYNC. The sparsecopy
code does not seem to do it?

On both host and guest every single write has 'S' set. You are correct that
sparsecopy doesn't open with O_SYNC. Perhaps it's the fact that it's the
fdatasync() which is causing the writes from the page cache (and that by
definition is blocking) which causes blktrace to see 'S' bits? I think it
was writing 4k blocks and syncing every 16k.

Alex Bligh

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.