Обсуждение: Server 9.1 doesn't restart and no output/log - data files seem corrupted

Hi,

The server was running correctly.
I didn't changed the configuration.
I did an apt-get upgrade and it didn't restarted.
The problem is maybe not related to the upgrade but appeared during the
restart.

I did the same upgrade on another (very similar) server and didn't have
any problem.
The configuration files and files permissions are the same.


The PostgreSQL packages that were upgraded:
   postgresql-9.1:amd64 (9.1.8-1, 9.1.9-0wheezy1)
   postgresql-client-9.1:amd64 (9.1.8-1, 9.1.9-0wheezy1)
   libpq-dev:amd64 (9.1.8-1, 9.1.9-0wheezy1)
   libpq5:amd64 (9.1.8-1, 9.1.9-0wheezy1)


The commands I tried to start the server:
# service postgresql start
[FAIL] Starting PostgreSQL 9.1 database server: main[....] The
PostgreSQL server failed to start. Please check the log output. ... failed!
  failed!

# pg_ctlcluster 9.1 main start
The PostgreSQL server failed to start. Please check the log output.


There's nothing in the log files.
I checked their permissions.
I renamed the files but they were not created again.


I created a new cluster and it started.
So I guess there's some corruption in the data files.


I replaced the data files with an old copy from 12 hours before but the
server won't start.


I tried to use pg_resetxlog:
# su postgres -c '/usr/lib/postgresql/9.1/bin/pg_resetxlog
/var/lib/postgresql/9.1/main'
The database server was not shut down cleanly.
Resetting the transaction log might cause data to be lost.
If you want to proceed anyway, use -f to force reset.

Then:
# su postgres -c '/usr/lib/postgresql/9.1/bin/pg_resetxlog -f
/var/lib/postgresql/9.1/main'
Transaction log reset

But it still won't start.


What can I do?

Thanks in advance.

I managed to solve the problem.

For the record, I did the following:

mkdir -p ~/postgresql/{etc,lib,log}
cp -a /etc/postgresql/9.1/main ~/postgresql/etc/old-install
cp -a /var/lib/postgresql/9.1/main ~/postgresql/lib/old-install
cp -a /var/log/postgresql ~/postgresql/log/old-install

apt-get remove --purge postgresql-9.1 postgresql-client-9.1
postgresql-client-common postgresql-common
rm -rf /etc/postgresql
rm -rf /var/lib/postgresql

apt-get install postgresql
service postgresql stop

cp -a /etc/postgresql/9.1/main ~/postgresql/etc/new-install
cp -a /var/lib/postgresql/9.1/main ~/postgresql/lib/new-install

rm -rf /var/lib/postgresql/9.1/main
cp -a ~/postgresql/lib/old-install /var/lib/postgresql/9.1/main
su postgres -c '/usr/lib/postgresql/9.1/bin/pg_resetxlog -f
/var/lib/postgresql/9.1/main'

I edited /etc/postgresql/9.1/main/postgresql.conf

service postgresql start

And it worked :-)



On 04/04/13 21:25, Pascal Polleunus wrote:
> Hi,
>
> The server was running correctly.
> I didn't changed the configuration.
> I did an apt-get upgrade and it didn't restarted.
> The problem is maybe not related to the upgrade but appeared during the
> restart.
>
> I did the same upgrade on another (very similar) server and didn't have
> any problem.
> The configuration files and files permissions are the same.
>
>
> The PostgreSQL packages that were upgraded:
>    postgresql-9.1:amd64 (9.1.8-1, 9.1.9-0wheezy1)
>    postgresql-client-9.1:amd64 (9.1.8-1, 9.1.9-0wheezy1)
>    libpq-dev:amd64 (9.1.8-1, 9.1.9-0wheezy1)
>    libpq5:amd64 (9.1.8-1, 9.1.9-0wheezy1)
>
>
> The commands I tried to start the server:
> # service postgresql start
> [FAIL] Starting PostgreSQL 9.1 database server: main[....] The
> PostgreSQL server failed to start. Please check the log output. ... failed!
>   failed!
>
> # pg_ctlcluster 9.1 main start
> The PostgreSQL server failed to start. Please check the log output.
>
>
> There's nothing in the log files.
> I checked their permissions.
> I renamed the files but they were not created again.
>
>
> I created a new cluster and it started.
> So I guess there's some corruption in the data files.
>
>
> I replaced the data files with an old copy from 12 hours before but the
> server won't start.
>
>
> I tried to use pg_resetxlog:
> # su postgres -c '/usr/lib/postgresql/9.1/bin/pg_resetxlog
> /var/lib/postgresql/9.1/main'
> The database server was not shut down cleanly.
> Resetting the transaction log might cause data to be lost.
> If you want to proceed anyway, use -f to force reset.
>
> Then:
> # su postgres -c '/usr/lib/postgresql/9.1/bin/pg_resetxlog -f
> /var/lib/postgresql/9.1/main'
> Transaction log reset
>
> But it still won't start.
>
>
> What can I do?
>
> Thanks in advance.
>
>
>

Thanks for your post. That helped me for solving my issue.



--
View this message in context:
http://postgresql.1045698.n5.nabble.com/Server-9-1-doesn-t-restart-and-no-output-log-data-files-seem-corrupted-tp5750824p5811563.html
Sent from the PostgreSQL - admin mailing list archive at Nabble.com.