Обсуждение: Restoring postgresql 9.4 crashed database

On Mon, Apr 11, 2016 at 7:29 AM, Babacar Gningue
<gningue.babacar.ita@gmail.com> wrote:
> Can anyone help me to restore my crashed database. Is there a way to use
> pg_resetxlog. I read a number of documentations telling about it but i'm
> always seeing the same log when i restart my postgresql service ?

> 2016-04-11 8:47:59 p.m. GMT [1710-1] LOG: the database system was
> interrupted during a restore at 2016-04-11 11:04:19 GMT
> 2016-04-11 8:47:59 p.m. GMT [1710-2] TIP: This probably means that the data
> has been corrupted and you
> need to use the last backup for recovery.
> 2016-04-11 8:47:59 p.m. GMT [1710-3] LOG: the database system was not shut
> properly; restoration
> automatic current
> 2016-04-11 8:47:59 p.m. GMT [1710-4] LOG: the re-execution starts at 4 /
> 32DF8E90
> 2016-04-11 8:47:59 p.m. GMT [1710-5] FATAL: could not access the transaction
> status 5598
> 2016-04-11 8:47:59 p.m. GMT [1710-6] DETAIL: Could not read the file
> "pg_multixact / members / 0000" at offset 57344: Success
> 2016-04-11 8:47:59 p.m. GMT [1710-7] BACKGROUND: xlog redo create mxiD 5598
> offset 11603 Nmembers 2: 6486476 (keysh) 6,486,482 (keysh)

You posted to a JDBC list, and this has nothing to do with JDBC.
You are much more likely to get useful responses by posting to
pgsql-general@postgresql.org.  When you do, please be very specific
about what version of the software you are using -- perhaps run one
of the executables with the --version option if you can't get the
server started to SELECT version();

Also, be very specific about how the backup was made -- scripts or
copy paste from the shell window or history are best.  This sort of
error could be caused by improper backup or restore procedures, or
by not using the latest minor release of the 9.4 major release --
there have been some important bug fixes related to multixact
handling.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On Mon, Apr 11, 2016 at 7:29 AM, Babacar Gningue
<gningue.babacar.ita@gmail.com> wrote:
> Can anyone help me to restore my crashed database. Is there a way to use
> pg_resetxlog. I read a number of documentations telling about it but i'm
> always seeing the same log when i restart my postgresql service ?

> 2016-04-11 8:47:59 p.m. GMT [1710-1] LOG: the database system was
> interrupted during a restore at 2016-04-11 11:04:19 GMT
> 2016-04-11 8:47:59 p.m. GMT [1710-2] TIP: This probably means that the data
> has been corrupted and you
> need to use the last backup for recovery.
> 2016-04-11 8:47:59 p.m. GMT [1710-3] LOG: the database system was not shut
> properly; restoration
> automatic current
> 2016-04-11 8:47:59 p.m. GMT [1710-4] LOG: the re-execution starts at 4 /
> 32DF8E90
> 2016-04-11 8:47:59 p.m. GMT [1710-5] FATAL: could not access the transaction
> status 5598
> 2016-04-11 8:47:59 p.m. GMT [1710-6] DETAIL: Could not read the file
> "pg_multixact / members / 0000" at offset 57344: Success
> 2016-04-11 8:47:59 p.m. GMT [1710-7] BACKGROUND: xlog redo create mxiD 5598
> offset 11603 Nmembers 2: 6486476 (keysh) 6,486,482 (keysh)

You posted to a JDBC list, and this has nothing to do with JDBC.
You are much more likely to get useful responses by posting to
pgsql-general@postgresql.org.  When you do, please be very specific
about what version of the software you are using -- perhaps run one
of the executables with the --version option if you can't get the
server started to SELECT version();

Also, be very specific about how the backup was made -- scripts or
copy paste from the shell window or history are best.  This sort of
error could be caused by improper backup or restore procedures, or
by not using the latest minor release of the 9.4 major release --
there have been some important bug fixes related to multixact
handling.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company