Обсуждение: Could not locate a valid checkpoint record

Hi all

I tried to start a PostgresSQL instance this morning, but it failed. So
I checked the log:

2010-06-24 09:00:28 LOG:  database system was shut down at 2010-06-24
08:13:58

2010-06-24 09:00:28 LOG:  record with zero length at 9/E0E9D200

2010-06-24 09:00:28 LOG:  invalid primary checkpoint record

2010-06-24 09:00:28 LOG:  record with zero length at 9/E0E9D1B0

2010-06-24 09:00:28 LOG:  invalid secondary checkpoint record

2010-06-24 09:00:28 PANIC:  could not locate a valid checkpoint record

This application has requested the Runtime to terminate it in an unusual
way.

Please contact the application's support team for more information.

2010-06-24 09:00:28 LOG:  startup process (PID 1644) exited with exit code 3

2010-06-24 09:00:28 LOG:  aborting startup due to startup process failure

2010-06-24 09:00:29 LOG:  logger shutting down

I know this error requires you to restore from backup, but,
unfortunatelly, we have no backup.
There is any way I can recover the data?

Thanks

Norberto

Norberto Delle <betodelle@gmail.com> writes:
> 2010-06-24 09:00:28 PANIC:  could not locate a valid checkpoint record

> I know this error requires you to restore from backup, but,
> unfortunatelly, we have no backup.
> There is any way I can recover the data?

pg_resetxlog should allow you to restart the database, but it's anyone's
guess as to whether the data will be corrupt.  A dump and reload would
be a good idea.  You also need to investigate what happened to your
pg_xlog files and take steps to make sure it doesn't happen again.
(Since this is evidently Windows, I would suggest being sure you're
using a recent/supported PG release, and take a close look at what AV
software you're using.)

            regards, tom lane

Em 24/6/2010 11:48, Tom Lane escreveu:
> Norberto Delle<betodelle@gmail.com>  writes:
>
>> 2010-06-24 09:00:28 PANIC:  could not locate a valid checkpoint record
>>
>
>> I know this error requires you to restore from backup, but,
>> unfortunatelly, we have no backup.
>> There is any way I can recover the data?
>>
> pg_resetxlog should allow you to restart the database, but it's anyone's
> guess as to whether the data will be corrupt.  A dump and reload would
> be a good idea.  You also need to investigate what happened to your
> pg_xlog files and take steps to make sure it doesn't happen again.
> (Since this is evidently Windows, I would suggest being sure you're
> using a recent/supported PG release, and take a close look at what AV
> software you're using.)
>
>             regards, tom lane
>
>
Yes, this instance is running on Windows XP.
I ran pg_resetxlog against the data directory and I'm reloading the
dumped data in
a new cluster i created.
The machine is running Avira AntVir, but Avira played well with Postgres
so far.
What kind of data verification do you recommend? A sucessfull
dump/reload is a good sign that the data
is healthy?

Thanks for the attention

Norberto