Re: FTP Troubles w/ many servers

Thanks for the response...and more power to you for keeping distribution
growth alive - I'm always looking for a new distro to dig into.  

To answer your question, no it seems to be pretty random about where it
dies.  Here's the very end of the debug output for "grab --update
--debug -W"...

Net::FTP=GLOB(0x8520a34)>>> TYPE I
Net::FTP=GLOB(0x8520a34)<<< 200 Switching to Binary mode.
Net::FTP=GLOB(0x8520a34)>>> PASV
Net::FTP=GLOB(0x8520a34)<<< 227 Entering Passive Mode
(216,254,0,38,45,219)
Net::FTP=GLOB(0x8520a34)>>> LIST
Net::FTP=GLOB(0x8520a34)<<< 150 Here comes the directory listing.
Net::FTP=GLOB(0x8520a34)<<< 226 Directory send OK.
 ->Indexed [18] RPMS from sourceforge@SPEAKEASY/glunarclock
 ->Indexed [18] RPMS from sourceforge@SPEAKEASY/glunarclock
   -Problem parsing packages from sourceforge@SPEAKEASY/glunarclock (use
verbose to view more details)
 WARN: Could not connect to speakeasy.rpmfind.net!
Can't call method "login" on an undefined value at /usr/bin/grab line
1715.
[root@localhost cube2003]#

Often, it quits on getting the directory listing, with a message message
about a timout in FTP.pm.  It just seems to depend on where the hangup
happens which message I get.  If all else fails, is there any way for me
to just update in chunks at a time so that I don't have to COMPLETELY
restart every time it does sieze up.

Thanks for your help and this cool little program!

Keller



On Mon, 2004-02-09 at 16:54, Greg Kurtzer wrote: 
> Does it always die in the same spot? Can you also send the debug output 
> where it is dying?
> 
> To comment on the list being dormant, I just posted a blurb to the GRAB
> homepage. Here is a snippet:
> 
> 
> NEWS: GRAB development has stopped!
> 
> I have not had the time to maintain GRAB since the RedHat announcement that
> they are dropping support of their free Operating Systems. What have I been
> doing you ask? Myself and many other community developers have grouped
> together to form a new distribution of Linux. This distribution can be
> described as a cross between Debian and RedHat (a community managed RPM based
> system). It is not Fedora!
> 
> The distribution is cAos (http://caosity.org) and development is moving very
> fast with the hopes to create a usable distribution for people to migrate
> their EOL'ed RedHat systems to.
> 
> If someone wishes to continue maintaining GRAB, please contact me and/or the
> mailing list.
> 
> If you wish to use GRAB, please let me refer you to YUM
> (http://linux.duke.edu/projects/yum/). 
> 
> 
> On Mon, Feb 09, 2004 at 03:31:01PM -0600, Keller told me:
> > I know that this list is fairly dormant, but I hope someone will get
> > this who can help me.  I'm currently trying to index a bunch of rpms
> > from one server in many directories (in my case, the different
> > directories in ftp://speakeasy.rpmfind.net/linux/).  If I try to just
> > recursively index that one directory, I get a timeout from grab.  If I
> > try to manually input each of the subdirectories (/mandrake, /freshmeat,
> > etc) of that directory, then grab runs fine for a while, but inevitably
> > gives me a "cannot connect to rpmfind.speakeasy.net" and then ends the
> > update.  If I run with debug, I find that it connects to the server fine
> > many times and then simply can't.  I assume that it's just an anomaly
> > that occurs once in a while, but since I'm updating so many directories,
> > occurs once in the many connections that grab makes for my server list. 
> > In case it helps, here's the command line I'm using ::
> > 
> > grab --update -W -debug
> > 
> > I've also attached my config and servers files, just in case they help. 
> > Is there any way to force grab to continue even after it can't connect
> > to a server?  Can I setup perl's ftp to retry a server if it's unable to
> > connect?  Thanks so much for the help!
> > 
> > Sincerely, 
> > Keller
> > 
> > 
> > _______________________________________________
> > Grab mailing list
> > [email protected]
> > http://www.runlevelzero.net/mailman/listinfo/grab