mkcol patch for lighttpd support

Justin Huff jjhuff at mspin.net
Mon Nov 19 21:44:07 UTC 2007


Ask,

We've been running pretty much this exact patch for about 6months in
production with no problems.  We had a similar issue and thought that
lighttpd was a great solution:)

--Justin

Ask Bjørn Hansen wrote:
> I'm storing some big (~100MB) files in one installation and perlbal is
> eating up gobs of memory (gigabytes some of them ended up on).   Rather
> than tracking that down I just changed to lighttpd.
> 
> lighttpd, correctly apparently, returns 405 when it won't create a
> directory that already exists.    Brad earlier -
> http://lists.danga.com/pipermail/mogilefs/2007-April/000886.html -
> implied that it's important to detect.  Why do we care at all?  If we
> can't make the directory it just won't work anyway, no?  Why does it
> matter that we're sending extra bogus requests in that case?   Or does
> it just matter that the directory actually got created and we need to know?
> 
> If the latter, wouldn't it be better to just do a PROPFIND or some such
> after MKCOL to check that the directory is there?
> 
> 
>   - ask
> 
> 
> --- lib/MogileFS/Device.pm    (revision 1129)
> +++ lib/MogileFS/Device.pm    (working copy)
> @@ -366,7 +366,7 @@
>      my $ans = <$sock>;
> 
>      # if they don't support this method, remember that
> -    if ($ans && $ans =~ m!HTTP/1\.[01] (400|405|501)!) {
> +    if ($ans && $ans =~ m!HTTP/1\.[01] (400|501)!) {
>          $self->{no_mkcol} = 1;
>          # TODO: move this into method on device, which propogates to
> parent
>          # and also receive from parent.  so all query workers share
> this knowledge
> 
> 
> 


More information about the mogilefs mailing list