Getting started, Bad file descriptor

[Brad Fitzpatrick]

FWIW, I've been going through all our code and testing it on various
Redhat machines.  It's all getting there... I don't intend for it to be
Debian-only.  :)


On Sat, 30 Jul 2005, Kevin Lewandowski wrote:

> Hi, all. I've got it working now. Since the servers I'm putting
> mogilefs on aren't live yet I gave up and switched to Debian (I want
> to learn that anyway). Now it's all working fine, including the perl
> client and recent python client.
>
> thanks!
>
> Kevin
>
> On Jul 29, 2005, at 9:47 AM, <j.ignacio.leon at gmail.com> wrote:
>
> > To use IO::AIO you have to install it (CPAN is probably the easiest
> > way, or look for Marc Lehmann's perl contributions). However the fact
> > that you get the same error with aio_mode set to none seems to
> > indicate you have some other issues.
> > Juan
> >
> >
> > On 7/28/05, Kevin Lewandowski <kevin at discogs.com> wrote:
> >
> >> Okay, I've installed Perlbal 1.3 (fyi, it seems to be missing
> >> CommandContext.pm, I had to get that out of CVS).
> >>
> >> With aio_mode set to none I get the same error. With aio_mode set to
> >> ioaio I get the error: "ERROR: IO::AIO not available"
> >>
> >> Does that mean I'm missing something? I do have Linux-AIO-1.72
> >> installed.
> >>
> >> thanks,
> >> Kevin
> >>
> >> On Jul 26, 2005, at 1:52 PM, Mark Smith wrote:
> >>
> >>
> >>>> Hi, Mark. I did that and still get the same result. thanks
> >>>>
> >>>>
> >>>
> >>> Okay.  We've just made a release (1.3) of Perlbal that has lots of
> >>> changes
> >>> in it.  Can you try going and getting that, then installing it?
> >>>
> >>> Once you've got it installed, try running it with aio_mode set to
> >>> 'ioaio'
> >>> or 'none' and see if either of those work now.  The new version has
> >>> a bunch
> >>> of changes to how we handle asynchronous IO.
> >>>
> >>> If that still doesn't work, then it'd be useful to do an strace
> >>> of the
> >>> mogstored and try to find where it's erroring in the call to open
> >>> ().  You
> >>> may have to start inserting debugging in the part of the code that
> >>> prints
> >>> the internal error message, and try to find out why it's screwing
> >>> up.
> >>>
> >>> If you're not familiar with the best way to go about doing these,
> >>> let me
> >>> know and I can provide more detailed information.  :)
> >>>
> >>>
> >>> --
> >>> Junior (aka Mark Smith)
> >>> junior at danga.com
> >>>
> >>> Software Engineer
> >>> Six Apart / Danga Interactive
> >>>
> >>>
> >>>
> >>
> >>
> >>
> >
> >
>
>