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

Re: [Bacula-devel] Multiple fd plugins in one fileset: Restore problems


I've just looked at the code and in at least one place, a loop over all the 
instances of plugins is missing, and it is right where your seg fault occurs, 
so that could well be the problem.  I'll try to work on that tonight.

See below ...

On Monday 29 September 2008 16:55:55 Bastian Friedrich wrote:
> Hi,
> On Monday 29 September 2008, Kern Sibbald wrote:
> > - the code is still under important (heavy) development, and you didn't
> > specify what version of Bacula you are using.
> SVN snapshot from last friday; I am just now upgrading to a current
> snapshot (rev. 7662).
> > - Although the plugin interface has been changing, I have not incremented
> > the version numbers, which means that if you do not build Bacula and the
> > plugin at the same time, there is bound to be some problems.
> bpipe plugin, director and file daemon (and all other code) are built from
> the same sources and were installed at the same point of time.
> > - The plugin is not at all going to understand __RESTORE_LOCATION=xxx
> > /path... So I am not sure how that is going to work.
> Whoops. That was some leftover debugging code.
> "__RESTORE_LOCATION__==%r /path..." gives the restore location as an
> environment variable to the receiving executable. This part works fine. The
> problems persist when removing that stuff.
> > - Though it is possible, I don't see any connection between this and the
> > ACL problem. The code is very different.
> Both problems seem to be located in filed/restore.c; I do have the
> impression that both problems have something to do with
> inconsistent/insufficient state transitions when the stream type changes. I
> may well be wrong.
> > - Error conditions concerning the plugin are still being worked on to
> > ensure that it is robust.
> Good :)
> > - There may well be a bug.  I'll add multiple Plugin calls to my test
> > case.
> Even better!
> > - If you can ensure you are on the latest SVN update, it would help, but
> > realize the plugin code is still changing on a daily basis as we are just
> > now developing a "serious" plugin, so we are correcting problems "real
> > time".
> ACK.
> > - If you are on a recent SVN update and you get a crash, I would like to
> > see all the job output (perhaps a debug -d 50 listing) and most important
> > a traceback.
> I suppose you are requesting a -d 50 from the fd; it is attached for cases
> (4) and (5), referring the cases (4) and (4), the second, of my first mail
> (mis-numbered them there).
> Traceback of case 5 attached as well.

Thanks that was a big help in pointing out a bug.

> > The above remarks are not at all designed to discourage you -- we *need*
> > testers.
> I did not feel offended. :)

I didn't imagine so, but some people seem to get discourged if I don't 
immediately give positive feedback, so I'm just trying to be careful.

Best regards,


> Thx & best regards
>    Bastian

This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
Bacula-devel mailing list

This mailing list archive is a service of Copilotco.