Thread: pg_pwd problem, 2nd try

If I use the ALTER USER sql statement, the file pg_pwd gets written in an
older plain text format & the user can no longer authenticate. However,
using pg_passwd from the command line reates a file similar to /etc/passwd
including the username and an encrypted password. Known problem?

Ken

"Ken J. Wright" <ken@ori-ind.com> writes:
> If I use the ALTER USER sql statement, the file pg_pwd gets written in an
> older plain text format & the user can no longer authenticate. However,
> using pg_passwd from the command line reates a file similar to /etc/passwd
> including the username and an encrypted password. Known problem?

It is now ;-)

Peter changed the backend's idea about the format of pg_pwd, but
I guess he forgot about pg_passwd ...

I wonder whether we even need pg_passwd in its present form anymore.
Maybe it should just issue an ALTER USER command?
        regards, tom lane

On 2000-02-22, Tom Lane mentioned:

> "Ken J. Wright" <ken@ori-ind.com> writes:
> > If I use the ALTER USER sql statement, the file pg_pwd gets written in an
> > older plain text format & the user can no longer authenticate. However,
> > using pg_passwd from the command line reates a file similar to /etc/passwd
> > including the username and an encrypted password. Known problem?

Uh, you should not use pg_passwd to change pg_pwd. It should only be used
to create auxilliary password files for pg_hba.conf lines like
.... passwd my_separate_password_file

> Peter changed the backend's idea about the format of pg_pwd, but

Not really. Only the unused fields are filled in with dummies.

> I guess he forgot about pg_passwd ...

Unrelated. See above.

> I wonder whether we even need pg_passwd in its present form anymore.
> Maybe it should just issue an ALTER USER command?

Nope.


-- 
Peter Eisentraut                  Sernanders väg 10:115
peter_e@gmx.net                   75262 Uppsala
http://yi.org/peter-e/            Sweden