webadmin-devel

Re: Bug in fsdump module...

[<jam...@te...>]

OK, replying to myself, but after I fixed the problem with setting the
confiuguration
I had a simular problem in when running it in execute_dump().  I used the
same strategy
to fix it.  As I was fixing it thouugh, I noticed that at the end after
running the dump command
the return code from dump was not checked (I would never do that (-;).
Anyway, this patch contains
the previous change, the new change for the same error in execute_dump, and
it contains a suggested
fix for not checking the return code.  I forgot to mention this is against
the 0.980 source with no updates.

Here is the  new patch:

*** redhat-linux-lib.pl Fri Jun 21 16:13:00 2002
--- redhat-linux-lib.pl.orig    Fri Jun 21 14:54:52 2002
***************
*** 190,239 ****
  # Executes a dump and displays the output
  sub execute_dump
  {
- my    $dump_data = shift;
  local $fh = $_[1];
  local ($cmd, $flag);
!
! #
! # Verify we were passed the correct args
! if (!ref($dump_data)) {
!     print "<B>Error Call to execute_dump() in redhat-linux-lib.pl with invalid arguments!</B>";
!     return 0
! }
!
! if ($dump_data->{'huser'}) {
!       $flag = " -f '" . $dump_data->{'huser'} . '@' . $dump_data->{'host'} .
!           ':' . $dump_data->{'hfile'} . "'";
        }
! elsif ($dump_data->{'host'}) {
!       $flag = " -f '" . $dump_data->{'host'} . ':' . $dump_data->{'hfile'} . "'";
        }
  else {
!       $flag = " -f '" . $dump_data->{'file'} . "'";
        }
! if ($dump_data->{'fs'} eq 'xfs') {
        # xfs backup
        $cmd = "xfsdump -l $_[0]->{'level'}";
        $cmd .= $flag;
!       $cmd .= " -L '" . $dump_data->{'label'} . "'" if ($dump_data->{'label'});
!       $cmd .= " -M '" . $dump_data->{'label'} . "'" if ($dump_data->{'label'});
!       $cmd .= " -z '" . $dump_data->{'max'} . "'"   if ($dump_data->{'max'});
!       $cmd .= " -A" if ($dump_data->{'noattribs'});
!       $cmd .= " -F" if ($dump_data->{'over'});
!       $cmd .= " -J" if ($dump_data->{'noinvent'});
!       $cmd .= " -o" if ($dump_data->{'overwrite'});
!       $cmd .= " -E -F" if ($dump_data->{'erase'});
!       $cmd .= " '" . $dump_data->{'dir'} . "'";
        }
  else {
        # ext2/3 backup
!       $cmd = "dump -" . $dump_data->{'level'};
        $cmd .= $flag;
!       $cmd .= " -u" if ($dump_data->{'update'});
!       $cmd .= " -M" if ($dump_data->{'multi'});
!       $cmd .= " -L '" . $dump_data->{'label'} . "'" if ($dump_data->{'label'});
!       $cmd .= " -B " . $dump_data->{'blocks'} if ($dump_data->{'blocks'});
!       $cmd .= " '" . $dump_data->{'dir'} . "'";
        }

  &system_logged("sync");
--- 190,229 ----
  # Executes a dump and displays the output
  sub execute_dump
  {
  local $fh = $_[1];
  local ($cmd, $flag);
! if ($_[0]->{'huser'}) {
!       $flag = " -f '$_[0]->{'huser'}@$_[0]->{'host'}:$_[0]->{'hfile'}'";
        }
! elsif ($_[0]->{'host'}) {
!       $flag = " -f '$_[0]->{'host'}:$_[0]->{'hfile'}'";
        }
  else {
!       $flag = " -f '$_[0]->{'file'}'";
        }
! if ($_[0]->{'fs'} eq 'xfs') {
        # xfs backup
        $cmd = "xfsdump -l $_[0]->{'level'}";
        $cmd .= $flag;
!       $cmd .= " -L '$_[0]->{'label'}'" if ($_[0]->{'label'});
!       $cmd .= " -M '$_[0]->{'label'}'" if ($_[0]->{'label'});
!       $cmd .= " -z '$_[0]->{'max'}'" if ($_[0]->{'max'});
!       $cmd .= " -A" if ($_[0]->{'noattribs'});
!       $cmd .= " -F" if ($_[0]->{'over'});
!       $cmd .= " -J" if ($_[0]->{'noinvent'});
!       $cmd .= " -o" if ($_[0]->{'overwrite'});
!       $cmd .= " -E -F" if ($_[0]->{'erase'});
!       $cmd .= " '$_[0]->{'dir'}'";
        }
  else {
        # ext2/3 backup
!       $cmd = "dump -$_[0]->{'level'}";
        $cmd .= $flag;
!       $cmd .= " -u" if ($_[0]->{'update'});
!       $cmd .= " -M" if ($_[0]->{'multi'});
!       $cmd .= " -L '$_[0]->{'label'}'" if ($_[0]->{'label'});
!       $cmd .= " -B $_[0]->{'blocks'}" if ($_[0]->{'blocks'});
!       $cmd .= " '$_[0]->{'dir'}'";
        }

  &system_logged("sync");
***************
*** 249,285 ****
                }
        }
  close(CMD);
- my $rc = $? >> 8;
- if($rc != 0) {
-     print <<EOF;
- <B>
- <P>ERROR Running dump!</P>
-
- COMMAND:  ${cmd}
-      RC:  ${rc}
- EOF
-     }
-
  }

  # dump_dest(&dump)
  sub dump_dest
  {
! my $dump_data = shift;
!
! if(!ref($dump_data)) {
!     return "<B>Error Call to dump_dest() in redhat-linux-lib.pl with invalid arguments!</B>";
! }
! if ($dump_data->{'file'}) {
!       return "<tt>".&html_escape($dump_data->{'file'})."</tt>";
        }
  elsif ($_[0]->{'huser'}) {
!       return "<tt>" . &html_escape($dump_data->{'huser'} . '@' .
!               $dumpdata->{'host'} . ':' . $dump_data->{'hfile'}) . "</tt>";
        }
  else {
!       return "<tt>" . &html_escape($dump_data->{'host'} . ':' .
!           $dump_data->{'hfile'}) . "</tt>";
        }
  }

--- 239,257 ----
                }
        }
  close(CMD);
  }

  # dump_dest(&dump)
  sub dump_dest
  {
! if ($_[0]->{'file'}) {
!       return "<tt>".&html_escape($_[0]->{'file'})."</tt>";
        }
  elsif ($_[0]->{'huser'}) {
!       return "<tt>".&html_escape("$_[0]->{'huser'}@$_[0]->{'host'}:$_[0]->{'hfile'}")."</tt>";
        }
  else {
!       return "<tt>".&html_escape("$_[0]->{'host'}:$_[0]->{'hfile'}")."</tt>";
        }
  }

Cheers...james

Re: Bug in fsdump module...

[<jam...@te...>]

Jamie,

I downloaded your filesystem update module, and it only has your fix in the
first funciton, and not
the second one, execute_dump().  And now completely understanding the
problem I see that it is
also in parse_restore().   Lastly, I really do think some code needs to be
at the bottome of execute_dump()
to check the return code of the dump command and let the user know if there
was an error.

Cheers...james

Password Module and PAM bug?

[<ra...@si...>]

  First, I would like to mention that the Password Module for Usermin can
include Solaris as a supported OS, I don't know if this was an oversight.
On the otherhand, if there is a bug that I am unaware of which affects
Solaris users, it would be nice to know.


  The other item is in reguards to PAM and authentication of users in
Usermin.  While recently searching a problem to which a user was
connecting via Usermin, but failed an FTP session, I identified that the
FTP session was failing because the password had expired.  There was no
notification in Usermin, however.  I am not certain at this time if this
is a bug/rfe for Usermin, or a bug/rfe in the PAM module, however, I
wanted to make folks on this list aware of this issue.  When I can put in
the cycles to hunt down what is needed, I will pass on the solution
(unless someone else knows the answer already).


rf

Re: Password Module and PAM bug?

[Jamie C. <jca...@we...>]

ra...@si... wrote:

> 
>   First, I would like to mention that the Password Module for Usermin can
> include Solaris as a supported OS, I don't know if this was an oversight.
> On the otherhand, if there is a bug that I am unaware of which affects
> Solaris users, it would be nice to know.

I've just tried it on my own solaris system, and it seems to work well.

In the next usermin release, the module will use PAM for password changing

on Solaris.


>   The other item is in reguards to PAM and authentication of users in
> Usermin.  While recently searching a problem to which a user was
> connecting via Usermin, but failed an FTP session, I identified that the
> FTP session was failing because the password had expired.  There was no
> notification in Usermin, however.  I am not certain at this time if this
> is a bug/rfe for Usermin, or a bug/rfe in the PAM module, however, I
> wanted to make folks on this list aware of this issue.  When I can put in
> the cycles to hunt down what is needed, I will pass on the solution
> (unless someone else knows the answer already).


Normally PAM should check if an account has expired, and if so deny authentication.
Usermin never does any such checking itself ..

  - Jamie

Re: Bug in fsdump module...

[Jamie C. <jca...@we...>]

jam...@te... wrote:

> Jamie,
> 
> I downloaded your filesystem update module, and it only has your fix in the
> first funciton, and not
> the second one, execute_dump().  And now completely understanding the
> problem I see that it is
> also in parse_restore().   Lastly, I really do think some code needs to be
> at the bottome of execute_dump()
> to check the return code of the dump command and let the user know if there
> was an error.


Actually, I found another occurance of the same error when adding code to
display a success/failure message, so I've released a new update that fixes
both problems.

  - Jamie