Thread: Killing the backends

Killing the backends

From
Helvio Hild
Date:
SELECT int8in(8);
     OVER psql do:
llamados=> SELECT int8in(8);
server closed the connection unexpectedly
        This probably means the server terminated abnormally
                before or while processing the request.
                The connection to the server was lost. Attempting reset:
Failed.
                !>



        In  psql superuser an other terminal:
        NOTICE:  Message from PostgreSQL backend:
                The Postmaster has informed me that some other backend
                        died abnormally and possibly corrupted shared
memory.
                                I have rolled back the current
transaction and am
                                        going to terminate your database
system connection and exit.
                                                Please reconnect to the
database system and repeat your query.


        In other Terminal
        BEFORE
:~$ pidof postmaster
        15489 15488 15474 338 337 335
        AFTER
 :~$ pidof postmaster
 338 337 335



PC:
i386->Celeron300
Linux->Debian->Kernel 2.2.18
Postgresql->7.2.1-2woody2 -> PostgreSQL 7.2.1 on i686-pc-linux-gnu,
compiled by GCC 2.95.4
                                                    BYE

---
[This E-mail scanned for viruses by Declude Virus]

Re: Killing the backends

From
Tom Lane
Date:
Helvio Hild <compulogia@coopcolorado.com.ar> writes:
> llamados=> SELECT int8in(8);
> server closed the connection unexpectedly

This is fixed as of PG 7.3.  It's not fixable in earlier releases
because they don't have a type-safe way of declaring I/O functions.

            regards, tom lane

Turkish locale bug

From
"Ismail Kizir"
Date:
Hi everybody,

Five years ago, i sent a patch for postgresql turkish locale patch, but
nobody has taken it into consideration.
Now, the source tree is totally different and pg team says they 'partially'
fixed turkish locale problem. But it is still non-functional.
I don't understand this, because the solution is very simple (at least, it
WAS simple for version 6.5... I don't know if the source of the problem is
the same now).

The source o the problem is :

In turkish alphabet, the lowercase of I is ý, not i.
And, postgresql parser uses a locale-dependent version of lowercase
function.
When it encounters an INSERT, it converts it to ýnsert, instead of insert...
And it gives an error.

The solution is very simple : Don't use a locale-dependent lowercase
function in the keyword parsing process.
Postgres team! Please, in the next minor release, fix this problem.
Here, in Turkey, we want to use Postgres, but your attitude discourages us.
If there is any subject that I may help, don't hesitate to write me... But
anyway please answer me.

Regards
Ismail Kizir

Re: Turkish locale bug

From
Devrim GUNDUZ
Date:
Hi,

On Wed, 5 May 2004, Ismail Kizir wrote:

> The solution is very simple : Don't use a locale-dependent lowercase
> function in the keyword parsing process.

Search the archives in -hackers -- with my name and Nicolai Tufar. This
discussion has been made already.

Also, there is a problem in glibc, for Turkish locale. Maybe this is
another cause of the problem.

--
Devrim GUNDUZ
devrim~gunduz.org                devrim.gunduz~linux.org.tr
            http://www.TDMSoft.com
            http://www.gunduz.org