Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

capture oracle pwd change in 3rd party application. help needed

0 views
Skip to first unread message

Lasher

unread,
Nov 5, 2003, 2:34:43 PM11/5/03
to
Hi,

I have clients using an application that allows users to change their
passwords. The application uses the 'ALTER USER xxx IDENTIFIED
BY.....' command. What I need to do is use Oracle to capture the
username and password and send the info to another Oracle instance on
a different server and update that users password.

Basically I need to keep the user's password in sync between two
different databases.

I also cannot change the application in anyway and therefore need to
do this from the Oracle side.

Any ideas would be great.........

Pete Finnigan

unread,
Nov 6, 2003, 5:48:26 AM11/6/03
to
Hi,

This all sounds a bit like hacking to me, not the best way to solve a
problem!, - maybe it would be better to ask the application
manufacturer?.

BUT, there are a number of possibilities that you could try, you could
implement a layer above OCI (it should be OCI unless its the thin
driver) and basically get the third party application to call your
version of OCI first, extract the string and then call the real OCI
layer. This is a hard way to crack the problem though. If it is java you
could do something illegal and decompile the java and modify it and
recompile - your supplier would not support you though!!!!

You don't have to access the username and password either in the
application, you could just run a process regularly that synchronises
password hashes between the databases using the "identified by values"
version of alter user. This would depend on how long you can wait for
passwords to synchronise. You could also use single sign on?

You could also grab the text off the wire or via trace, my recent paper
called "detecting SQL injection in Oracle" might help. You can get it at
http://www.petefinnigan.com/orasec.htm. Finally you might be able to
find a way to screen scrape the application at the terminal level -
there are a few commercial products to do this, mainly used for
automated testing.

hth
kind regards

Pete

PS: please don't cross post, most people read all of these groups!
--
Pete Finnigan
email:pe...@petefinnigan.com
Web site: http://www.petefinnigan.com - Oracle security audit specialists
Book:Oracle security step-by-step Guide - see http://store.sans.org for details.

Daniel Morgan

unread,
Nov 6, 2003, 2:46:08 PM11/6/03
to
Lasher wrote:

Go to $ORACLE_HOME/rdmbs/admin
Look at the file utlpwdmg.sql

If you have any business doing this you will be able to fill in the rest
of the picture.

Personally I agree with Pete. This is nonsense and worse than nonsense a
huge violation
of any reasonable definition of system security. The OEM should fix the
problem. And
my advise to you would be not to do this. That it can be done doesn't
mean that it should
be done. The entire idea stinks.

--
Daniel Morgan
http://www.outreach.washington.edu/ext/certificates/oad/oad_crs.asp
http://www.outreach.washington.edu/ext/certificates/aoa/aoa_crs.asp
damo...@x.washington.edu
(replace 'x' with a 'u' to reply)

Joel Garry

unread,
Nov 6, 2003, 5:55:54 PM11/6/03
to
Daniel Morgan <damo...@x.washington.edu> wrote in message news:<1068147990.413220@yasure>...

> Lasher wrote:
>
> >Hi,
> >
> >I have clients using an application that allows users to change their
> >passwords. The application uses the 'ALTER USER xxx IDENTIFIED
> >BY.....' command. What I need to do is use Oracle to capture the
> >username and password and send the info to another Oracle instance on
> >a different server and update that users password.
> >
> >Basically I need to keep the user's password in sync between two
> >different databases.
> >
> >I also cannot change the application in anyway and therefore need to
> >do this from the Oracle side.
> >
> >Any ideas would be great.........
> >
> >
> Go to $ORACLE_HOME/rdmbs/admin
> Look at the file utlpwdmg.sql
>
> If you have any business doing this you will be able to fill in the rest
> of the picture.
>
> Personally I agree with Pete. This is nonsense and worse than nonsense a
> huge violation
> of any reasonable definition of system security. The OEM should fix the
> problem. And
> my advise to you would be not to do this. That it can be done doesn't
> mean that it should
> be done. The entire idea stinks.

I'm not sure what is so wrong about this, at least using Pete's
suggestion of Identified by Values in a non-public environment? It
seems as reasonable as, say, copying /etc/passwd (or shadow
equivalents) and user files to synchronize users on two identical
servers.

jg
--
@home.com is bogus.
http://www.signonsandiego.com/news/metro/20031106-9999_2m6wage.html

Daniel Morgan

unread,
Nov 7, 2003, 1:21:09 AM11/7/03
to
My objection is that it would take me a matter of minutes to make myself an account on another
machine on which I had no permissions. It is a hacker's delight.

Pete Finnigan

unread,
Nov 7, 2003, 1:34:24 PM11/7/03
to
>> My objection is that it would take me a matter of minutes to
> make myself an account on another
> machine on which I had no permissions. It is a hacker's delight.

Hi Daniel,

I think there is another point to make here is that we are not
implementing this but just discussing possible solutions without knowing
the application or architecture, tools, requirements etc.... I would say
that a script to synchronise password hash values should be run in a
secure manner and also would not add new accounts, just synchronise old
ones. I would also re-iterate this isn't the way to fix an issue like
this, why does this application need to have synchronised access to two
databases? and why isn't the manufacturer involved.

kind regards

Pete

Daniel Morgan

unread,
Nov 7, 2003, 5:50:45 PM11/7/03
to
Pete Finnigan wrote:
        My objection is that it would take me a matter of minutes to 
      
   make myself an account on another 
   machine on which I had no permissions. It is a hacker's delight.
    
Hi Daniel,

I think there is another point to make here is that we are not
implementing this but just discussing possible solutions without knowing
the application or architecture, tools, requirements etc.... I would say
that a script to synchronise password hash values should be run in a
secure manner and also would not add new accounts, just synchronise old
ones. I would also re-iterate this isn't the way to fix an issue like
this, why does this application need to have synchronised access to two
databases? and why isn't the manufacturer involved. 

kind regards

Pete
  
My personal opinion? The person asking the question is trying to crack a database.
I've never seen an application with this architecture in 34 years in the business.

I'd really like to be wrong.

Anurag Varma

unread,
Nov 8, 2003, 12:52:39 AM11/8/03
to
 
"Daniel Morgan" <damo...@x.washington.edu> wrote in message news:1068245466.11957@yasure...
--------------
Actually one of the databases I manage runs on an application which does
something similar (Not the synchronization .. but the way it creates application accounts ...
by creating an Oracle account). The application being Bladerunner.
If you ever get a chance .. run (really fast) away from it.
:0)
Anurag
 

Pete Finnigan

unread,
Nov 8, 2003, 6:39:45 AM11/8/03
to
Thanks for the tip Anurag!
cheers

Daniel Morgan

unread,
Nov 8, 2003, 12:10:10 PM11/8/03
to
I'll do that.

BTW: I've tried to email you off-line and failed. Please send me your actual email address off-line. Thanks.

Anurag Varma

unread,
Nov 8, 2003, 12:45:10 PM11/8/03
to
Daniel,
 
I think I did reply you offline .. but not sure if the mail reached you.
 
Anyway, I realized that the email I was using to mail from this location
was wrong.
 
My email is: av...@hotmail.com
 
Anurag
"Daniel Morgan" <damo...@x.washington.edu> wrote in message news:1068311431.525724@yasure...

Daniel Morgan

unread,
Nov 8, 2003, 12:46:44 PM11/8/03
to
Thanks. I'll get back to you off-line.

Joel Garry

unread,
Nov 10, 2003, 5:59:42 PM11/10/03
to
Daniel Morgan <damo...@x.washington.edu> wrote in message news:<1068245466.11957@yasure>...

> Pete Finnigan wrote:
>
> >>> My objection is that it would take me a matter of minutes to
> >>>
> >>>
> >> make myself an account on another
> >> machine on which I had no permissions. It is a hacker's delight.
> >>
> >>
> >
> >Hi Daniel,
> >
> >I think there is another point to make here is that we are not
> >implementing this but just discussing possible solutions without knowing
> >the application or architecture, tools, requirements etc.... I would say
> >that a script to synchronise password hash values should be run in a
> >secure manner and also would not add new accounts, just synchronise old
> >ones. I would also re-iterate this isn't the way to fix an issue like
> >this, why does this application need to have synchronised access to two
> >databases? and why isn't the manufacturer involved.
> >
> >kind regards
> >
> >Pete
> >
> >
> My personal opinion? The person asking the question is trying to crack a
> database.
> I've never seen an application with this architecture in 34 years in the
> business.

Well, what do you think of SSO in Portal? The whole idea is to spread
a single password among apps. Then they give code to show the
passwords to admins! (ie, metalink note 205984.1). And you wind up
with two passwords, one of which is used in some places and the other
in others (such as whether you make all the, ahem, required public
synonyms through portal or sqlplus). IDENTIFIED GLOBALLY requires
some sort of syncronization between db's, and/or careful use of schema
independence.

>
> I'd really like to be wrong.

You may well be right, but By Values has been common knowledge for
generations. A couple of times I've almost written things like the OP
asked for, but it always turned out to be not necessary due to the
methods of copying the db. It's easy to envision an architecture like
that, though, particularly with unique ETL requirements. I've seen
worse - like admins keeping passwords in email so they know what to
change them all to manually.

jg
--
@home.com is bogus.

http://cbs.marketwatch.com/news/story.asp?guid=%7B954AA053-F953-43F3-BBC8-63D351A3BF2A%7D&siteid=google&dist=google

0 new messages