Thread: stopping "wal"

stopping "wal"

From
Manuel Trujillo
Date:
Hi.

How can I stop the recover process of wal? Well, first I will explain
this:
When more or two or three processes are working, and I stop the
database, if I have any problem stopping the DB, I kill the backends (is
a tester machine).
When I start the process "postmaster", the wal make a "recover of
processes", returning to the last situation.
I would like to know how to "say" to postmaster don't make a "recover"
from the wal files.

Sorry for this explanation, but my english is very poor... :(
Thank you very much.

Have a nice day ;-)
TooManySecrets

--
Manuel Trujillo         manueltrujillo@dorna.es
Technical Engineer      http://www.motograndprix.com
Dorna Sports S.L.       +34 93 4702864


Re: stopping "wal"

From
Tom Lane
Date:
Manuel Trujillo <manueltrujillo@dorna.es> writes:
> When more or two or three processes are working, and I stop the
> database, if I have any problem stopping the DB, I kill the backends (is
> a tester machine).

This is a bad idea.

> When I start the process "postmaster", the wal make a "recover of
> processes", returning to the last situation.
> I would like to know how to "say" to postmaster don't make a "recover"
> from the wal files.

And this would be worse.

If you don't care about your data, you can "initdb" to get rid of the
messed-up database.  If you do care, you shouldn't be doing any of the
above.

            regards, tom lane

Re: stopping "wal"

From
Manuel Trujillo
Date:
On Tue, 2002-02-12 at 16:16, Tom Lane wrote:

> And this would be worse.
>
> If you don't care about your data, you can "initdb" to get rid of the
> messed-up database.  If you do care, you shouldn't be doing any of the
> above.

The problem here, is a lot of programmers "start-stopping" the database
continuosly... :( And I can't make anything, because they need total
control over the database.
Here it is a bit of log after start the database in her actually state:

FATAL 1:  The database system is starting up
DEBUG:  proc_exit(0)
DEBUG:  shmem_exit(0)
DEBUG:  exit(0)
DEBUG:  BackendStartup: forked pid=18416 socket=8
DEBUG:  reaping dead processes
DEBUG:  child process (pid 18416) exited with exit
code 0
DEBUG:  INSERT @ 0/48F85360: prev 0/48F85320; xprev
 0/0; xid 0: XLOG - checkpoint: redo 0/48F85360; undo 0/0; sui 19; xid
2482396;
oid 3736683; shutdown
DEBUG:  XLogFlush: request 0/48F853A0; write 0/48F8
5360; flush 0/48F85360

DEBUG:  database system is ready
DEBUG:  proc_exit(0)
DEBUG:  shmem_exit(0)
DEBUG:  exit(0)
DEBUG:  reaping dead processes
DEBUG:  BackendStartup: forked pid=18422 socket=8
DEBUG:  connection: host=192.168.141.208 user=xxx
database=example

And make this, to 20 backend processes!!.

Tom say me to run "initdb", but... If possible to clean this wal's???

Thank you very much from here to Tom and the rest of pg-admin people.

Regards, Manuel.

--
Manuel Trujillo         manueltrujillo@dorna.es
Technical Engineer      http://www.motograndprix.com
Dorna Sports S.L.       +34 93 4702864