[OpenAFS] sqlite on AFS will not work, even with whole-file locking
Wed, 21 Apr 2010 13:28:20 +0200
On Mon, Apr 12, 2010 at 12:34:23AM -0400, Derrick Brashear wrote:
> On Sun, Apr 11, 2010 at 11:13 PM, Adam Megacz <firstname.lastname@example.org> wrote:
> > Brandon Simmons <email@example.com> writes:
> >> Thanks for the response. It seems like whole-file locking in sqlite
> >> would be a good choice for me in any case,
> >> In a situation where the whole-file locking scheme is used, would AF=
> >> be an acceptable choice? Would it be better than NFS?
> > I had the same idea, and tried it. =A0It does not work. =A0Your datab=
> > will get corrupted. =A0I never figured out why, although I did confir=
> > that sqlite was in fact requesting only whole-file locks.
> > It would be nice if it worked, though. =A0There are a lot of applicat=
> > out there where writes to the database are extremely rare, so
> > invalidating all the clients' caches is not a problem.
> do you happen to know what the corruption looked like (blocks of
> zeroes, just not readable, something else)
On 27 Oct 2008 I had a question about flock on AFS, because I did not
understand how flock should work on AFS at all. May I repeat this questio=
today I am totally confused how the flock(2) call should work on
Normally locking works in the following way:
1 fd =3D open("afs-file",O_RDWR)
do something with "afs-file"
When there are two processes (on different machines) executing that
code, the (2) flock call has to update the local copy of the afs-file,
otherwise locking is useless. And the (3) flock call has to sync the
local copy with the fileserver.
Writing a small test program I see that this synchronization isn't done.
How can I use the flock(2) call on AFS files?
Thank you for any help,
Hans-Werner Paulsen hans@MPA-Garching.MPG.DE
MPI f=FCr Astrophysik Tel 089-30000-2602
Karl-Schwarzschild-Str. 1 Fax 089-30000-2235=09