Date: Sun, 1 Dec 91 14:49:34 -0800
From: ambdata@dpi.inpe.br (Banco de Dados Ambiental)
Subject: Lock problems?
To: postgres@postgres.berkeley.edu
Sender: pg_adm@postgres.berkeley.edu
To: postgres@postgres.berkeley.edu
Message-Id: <9112012348.AA02627@dpi.inpe.br>

Hi,

        I have two postgres users running copies of a libpq aplication, so that
   they access the same database classes. Sometimes core is dumped, while trying
    to complete a transaction, after messages from the backend like this:

-------------------------------------------------------------------------------
beginWARN:Nov 29 15:32:45:cannot write block 0 of  [] blind


retrieve portal projects (Project.label) where Project.label = "NONE"NOTICE:Nov
   29 15:32:45:AbortTransaction and not in in-progress state

WARN:Nov 29 15:32:45:cannot write block 0 of  [] blind


fetch all in projectsNOTICE:Nov 29 15:32:45:AbortTransaction and not in in-progr
   ess state

WARN:Nov 29 15:32:45:cannot write block 0 of  [] blind

Segmentation fault (core dumped)
-------------------------------------------------------------------------------

        Sometimes no messages are sent. The same transaction is completed succes
   sfully as above:

-------------------------------------------------------------------------------
begin
retrieve portal projects (Project.label) where Project.label = "NONE"
fetch all in projects
close projects
append Project (label = "NONE" ,x1 = "4.390000e+04" ,y1  = "4.190000e+04" ,x2  =
    "6.410000e+04" ,y2  = "6.210000e+04" ,projection  = 1)
end
-------------------------------------------------------------------------------

        There is yet another message from the backend:

-------------------------------------------------------------------------------
NOTICE:Nov 29 14:56:32:I have been signalled by the postmaster.

NOTICE:Nov 29 14:56:32:Some backend process has died unexpectedly and possibly

NOTICE:Nov 29 14:56:32:corrupted shared memory.  The current transaction was

NOTICE:Nov 29 14:56:32:aborted, and I am going to exit.  Please resend the

NOTICE:Nov 29 14:56:32:last query. -- The postgres backend

Error: No response from the backend, exiting...
-------------------------------------------------------------------------------

Those situations seems to occur while two libpq applications are sharing tables
   in the same database. It seems like something related to the lock system...

Could someone point me any suggestion on how to manage this problem?

Thanks in advance.

        Joao Pedro.