LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: mysql not parsed in Ldirectord config.

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: mysql not parsed in Ldirectord config.
Cc: linux-ha-dev@xxxxxxxxxxxxxxxxxx
From: "Graham David Purcocks M.A.(Oxon.)" <grahamp@xxxxxxxxxxxxx>
Date: Thu, 12 May 2005 10:22:49 +0100
Perhaps Net::FTP changed underneath you? But you would expect it to use
syswrite from the outset. It is strange, but you definately can't
syswrite to a memory file in perl 5.8. Maybe its the newer versions of
perl?

Who knows

On Thu, 2005-05-12 at 07:33, Horms wrote:
> On Wed, May 11, 2005 at 11:09:23AM -0400, Purcocks, Graham wrote:
> > Horm
> > 
> > Here is a patch, I changed from a memory file to a virtual temporary
> > file. This is after appying both the mysql patch and the Postgres patch.
> > 
> > -----------------------------------------------------
> > --- /usr/sbin/ldirectord        2005-04-28 17:13:40.000000000 +0000
> > +++ ldirectord  2005-05-11 14:51:57.000000000 +0000
> > @@ -2079,7 +2079,8 @@
> > 
> >         &ld_debug(2, "Checking ftp server=$$r{server} port=$port");
> > 
> > -       open(MEMORY,'>', \$memory);
> > +       # Open temporary file
> > +       open(TMP,'>', undef);
> > 
> >         unless ($ftp = Net::FTP->new("$$r{server}:$port",
> >                         Timeout=>$$v{negotiatetimeout})) {
> > @@ -2090,10 +2091,14 @@
> >         $ftp->cwd("/");
> >         $ftp->binary();
> >         $ftp->pasv();
> > -       $ftp->get("$$r{request}", *MEMORY);
> > +       $ftp->get("$$r{request}", *TMP);
> >         $ftp->quit();
> > 
> > -       close(MEMORY);
> > +       seek TMP, 0, 0;
> > +       select TMP;
> > +       undef $/;
> > +       $memory = <TMP>;
> > +       close TMP;
> > 
> >         if ($memory =~ /$$r{receive}/) {
> >                 service_set($v, $r, "up");
> > --------------------------------------------------------- 
> > 
> > Is this OK?
> 
> That seems fine, I will go ahead and put it into CVS.
> Bit weird that the memory FD broke at some stage.
-- 

<Prev in Thread] Current Thread [Next in Thread>