Thread: URGENT: pg_dump & Postgres 7.2b4

URGENT: pg_dump & Postgres 7.2b4

From
"Christian Meunier"
Date:
Hi all, when i try to do a simple pg_dump i got the following errors:

2002-01-10 14:57:11 FATAL 2:  open of /usr/local/pgsql/data/pg_clog/0000
failed: No such file or directory
2002-01-10 14:57:11 DEBUG:  server process (pid 1201) exited with exit code
2
2002-01-10 14:57:11 DEBUG:  terminating any other active server processes
2002-01-10 14:57:11 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.
2002-01-10 14:57:11 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.
2002-01-10 14:57:11 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.
2002-01-10 14:57:11 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.
2002-01-10 14:57:11 DEBUG:  all server processes terminated; reinitializing
shared memory and semaphores
2002-01-10 14:57:11 DEBUG:  database system was interrupted at 2002-01-10
14:55:48 GMT
2002-01-10 14:57:11 DEBUG:  checkpoint record is at 0/1AD7A188
2002-01-10 14:57:11 DEBUG:  redo record is at 0/1AD75E88; undo record is at
0/0; shutdown FALSE
2002-01-10 14:57:11 DEBUG:  next transaction id: 2550628; next oid: 294388
2002-01-10 14:57:11 DEBUG:  database system was not properly shut down;
automatic recovery in progress
2002-01-10 14:57:11 DEBUG:  redo starts at 0/1AD75E88
2002-01-10 14:57:11 DEBUG:  ReadRecord: record with zero length at
0/1ADD14A8
2002-01-10 14:57:11 DEBUG:  redo done at 0/1ADD1484
2002-01-10 14:57:12 FATAL 1:  The database system is starting up
2002-01-10 14:57:12 FATAL 1:  The database system is starting up
2002-01-10 14:57:13 FATAL 1:  The database system is starting up
2002-01-10 14:57:13 FATAL 1:  The database system is starting up
2002-01-10 14:57:13 FATAL 1:  The database system is starting up
2002-01-10 14:57:13 FATAL 1:  The database system is starting up
2002-01-10 14:57:13 FATAL 1:  The database system is starting up
2002-01-10 14:57:13 DEBUG:  database system is ready
2002-01-10 14:57:14 ERROR:  Cannot insert a duplicate key into unique index
eq_profile_pkey
2002-01-10 14:57:36 ERROR:  Cannot insert a duplicate key into unique index
eq_profile_pkey
2002-01-10 14:57:57 ERROR:  Cannot insert a duplicate key into unique index
account_pkey
2002-01-10 14:58:18 ERROR:  Cannot insert a duplicate key into unique index
eq_profile_pkey
2002-01-10 14:58:44 ERROR:  Cannot insert a duplicate key into unique index
eq_profile_pkey

And in the console i got:

pg_dump: query to get data of sequence "account_num_seq" failed: FATAL 2:
open of /usr/local/pgsql/data/pg_clog/0000 failed: No such file or directory
server closed the connection unexpectedly
        This probably means the server terminated abnormally
        before or while processing the request.

The only file i got in directory pg_clog is : 0002

Help plz.
Thx in advance.




Re: URGENT: pg_dump & Postgres 7.2b4

From
"Christian Meunier"
Date:
this error also appears during normal operations (querying/updating
database).

Is there a way to clean up those commit log ?, i tried to remove 0002 but it
didnt help( it  was then doing error on 0002 too)

"Christian Meunier" <jelan@magelo.com> wrote in message
news:a1kc2b$2m6h$1@news.tht.net...
> Hi all, when i try to do a simple pg_dump i got the following errors:
>
> 2002-01-10 14:57:11 FATAL 2:  open of /usr/local/pgsql/data/pg_clog/0000
> failed: No such file or directory
> 2002-01-10 14:57:11 DEBUG:  server process (pid 1201) exited with exit
code
> 2
> 2002-01-10 14:57:11 DEBUG:  terminating any other active server processes
> 2002-01-10 14:57:11 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.
> 2002-01-10 14:57:11 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.
> 2002-01-10 14:57:11 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.
> 2002-01-10 14:57:11 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.
> 2002-01-10 14:57:11 DEBUG:  all server processes terminated;
reinitializing
> shared memory and semaphores
> 2002-01-10 14:57:11 DEBUG:  database system was interrupted at 2002-01-10
> 14:55:48 GMT
> 2002-01-10 14:57:11 DEBUG:  checkpoint record is at 0/1AD7A188
> 2002-01-10 14:57:11 DEBUG:  redo record is at 0/1AD75E88; undo record is
at
> 0/0; shutdown FALSE
> 2002-01-10 14:57:11 DEBUG:  next transaction id: 2550628; next oid: 294388
> 2002-01-10 14:57:11 DEBUG:  database system was not properly shut down;
> automatic recovery in progress
> 2002-01-10 14:57:11 DEBUG:  redo starts at 0/1AD75E88
> 2002-01-10 14:57:11 DEBUG:  ReadRecord: record with zero length at
> 0/1ADD14A8
> 2002-01-10 14:57:11 DEBUG:  redo done at 0/1ADD1484
> 2002-01-10 14:57:12 FATAL 1:  The database system is starting up
> 2002-01-10 14:57:12 FATAL 1:  The database system is starting up
> 2002-01-10 14:57:13 FATAL 1:  The database system is starting up
> 2002-01-10 14:57:13 FATAL 1:  The database system is starting up
> 2002-01-10 14:57:13 FATAL 1:  The database system is starting up
> 2002-01-10 14:57:13 FATAL 1:  The database system is starting up
> 2002-01-10 14:57:13 FATAL 1:  The database system is starting up
> 2002-01-10 14:57:13 DEBUG:  database system is ready
> 2002-01-10 14:57:14 ERROR:  Cannot insert a duplicate key into unique
index
> eq_profile_pkey
> 2002-01-10 14:57:36 ERROR:  Cannot insert a duplicate key into unique
index
> eq_profile_pkey
> 2002-01-10 14:57:57 ERROR:  Cannot insert a duplicate key into unique
index
> account_pkey
> 2002-01-10 14:58:18 ERROR:  Cannot insert a duplicate key into unique
index
> eq_profile_pkey
> 2002-01-10 14:58:44 ERROR:  Cannot insert a duplicate key into unique
index
> eq_profile_pkey
>
> And in the console i got:
>
> pg_dump: query to get data of sequence "account_num_seq" failed: FATAL 2:
> open of /usr/local/pgsql/data/pg_clog/0000 failed: No such file or
directory
> server closed the connection unexpectedly
>         This probably means the server terminated abnormally
>         before or while processing the request.
>
> The only file i got in directory pg_clog is : 0002
>
> Help plz.
> Thx in advance.
>
>
>



Re: URGENT: pg_dump & Postgres 7.2b4

From
Tom Lane
Date:
"Christian Meunier" <jelan@magelo.com> writes:
> Is there a way to clean up those commit log ?, i tried to remove 0002 but it
> didnt help( it  was then doing error on 0002 too)

You did *what*?

You may as well have done "rm -rf $PGDATA".  You're hosed.

I would have liked to look into why it was still trying to reference the
0000 segment after removing it; that suggests a logic problem somewhere.
But with the database now completely nonfunctional due to loss of the
active clog segment, there's probably no way to learn anything useful.

Don't remove files when you don't know what they are.

            regards, tom lane

Re: URGENT: pg_dump & Postgres 7.2b4

From
"Jelan"
Date:
Dont panic, it was just a stupid try and i had saved the file and then
restored it ;)

If you could tell me how to figure out why Postgres is still trying to
access a segment it deleted, i ll do it and report it.

Thanks in advance
----- Original Message -----
From: "Tom Lane" <tgl@sss.pgh.pa.us>
To: "Christian Meunier" <jelan@magelo.com>
Cc: <pgsql-general@postgresql.org>
Sent: Thursday, January 10, 2002 8:01 PM
Subject: Re: [GENERAL] URGENT: pg_dump & Postgres 7.2b4


> "Christian Meunier" <jelan@magelo.com> writes:
> > Is there a way to clean up those commit log ?, i tried to remove 0002
but it
> > didnt help( it  was then doing error on 0002 too)
>
> You did *what*?
>
> You may as well have done "rm -rf $PGDATA".  You're hosed.
>
> I would have liked to look into why it was still trying to reference the
> 0000 segment after removing it; that suggests a logic problem somewhere.
> But with the database now completely nonfunctional due to loss of the
> active clog segment, there's probably no way to learn anything useful.
>
> Don't remove files when you don't know what they are.
>
> regards, tom lane
>


Re: URGENT: pg_dump & Postgres 7.2b4

From
Tom Lane
Date:
"Jelan" <jelan@magelo.com> writes:
> Dont panic, it was just a stupid try and i had saved the file and then
> restored it ;)

Oh good.  In that case see my followup to pghackers.  I think your
problem is isolated to the sequence object(s), and can be summarized
as: nextval() will work, pg_dump'ing a sequence will not.  So, to get
out of trouble, try this: for each sequence, do nextval() to note the
current value, drop the sequence, recreate it, do setval() to restore
the count.  Then you'll be in a state where you can pg_dump.  Then try
not to go a million transactions between sequence creations and pg_dumps
while you're using 7.2b4 :-(.

There will be a fix in the next beta version.

Oh, and many thanks for finding this!  Would've been embarrassing to
have this glitch escape beta testing...

            regards, tom lane

Re: URGENT: pg_dump & Postgres 7.2b4

From
"Jelan"
Date:
I am not so sure that nextval() will work always, i tend to have some SQL
errors like:

Exception:ERROR:  Cannot insert a duplicate key into unique index
account_pkey

Something that i didnt have at all with 7.1

Best regards