On Monday 01 March 2010 @ 17:26, Tom Lane wrote:
> "Ed L." <pgsql(at)bluepolka(dot)net> writes:
> > On Monday 01 March 2010 @ 17:18, Tom Lane wrote:
> >> "Ed L." <pgsql(at)bluepolka(dot)net> writes:
> >>> There is one, looks like a typo got in.  How do I fix it?
> >>
> >> Hmm, a user named @, or what?
> >
> > Yes, a bogus user:
> >
> > "@" "" ""
> 
> Mph.  We really ought to fix things so that a quoted @ doesn't
>  get taken as an include file reference.  Odd that it's never
>  come up before.
> 
> Anyway, if you still have any open superuser sessions, the
>  best thing would be an ALTER USER RENAME.  If you don't,
>  you'll have to resort to manually editing the pg_auth file,
>  and then rename the user as soon as you can get in.

Killed the stuck postmaster with sigkill, edited the file, 
restarted postmaster, and it re-wrote the file with the bogus 
entry.  I don't have any superuser sessions open.  Is there 
another route?

Ed

In response to

• Re: Hung postmaster (8.3.9) at 2010-03-02 00:26:30 from Tom Lane

Responses

• Re: Hung postmaster (8.3.9) at 2010-03-02 00:49:44 from Ed L.
• Re: Hung postmaster (8.3.9) at 2010-03-02 00:58:49 from Tom Lane

pgsql-hackers by date

Next:	From: Greg Stark	Date: 2010-03-02 00:46:22
	Subject: Re: Hot Standby query cancellation and Streaming Replication integration
Previous:	From: Ed L.	Date: 2010-03-02 00:27:39
	Subject: Re: Hung postmaster (8.3.9)

pgsql-general by date

Next:	From: Ed L.	Date: 2010-03-02 00:49:44
	Subject: Re: Hung postmaster (8.3.9)
Previous:	From: Ed L.	Date: 2010-03-02 00:27:39
	Subject: Re: Hung postmaster (8.3.9)