Thread: initdb crashing, signal 5, etc

initdb crashing, signal 5, etc

From
Jacob Hanson
Date:
Had success with a snapshot from last week. Today's snapshot crashes
several times during initdb. As soon as "selecting default
max_connections" appears, postgres.exe crashes ("postgres.exe has
encountered a problem and needs to close."). After closing 6 of those
"10" appears and it continues to "selecting default shared_buffers".
11 crashes then "50" appears. It continues to the "No such file or
directory" error and postgres.exe crashes once, then the signal 5
message appears. Full initdb output is below.

Observations: During the crashes, postgresql.conf is created, but
remains 0 bytes. Cacls reports my "postgres" user has full access to
the entire c:\dev\pgsql tree, including the data dir, postgresql.conf
(after it has been created), and c:\dev\pgsql\share\timezone, which
certainly exists (note: NOT "/shared/timezone" necessarily), etc.






C:\dev\pgsql\data>initdb -d -L "c:/dev/pgsql/share"
Running in debug mode.
VERSION=7.5devel
PGDATA=C:/dev/pgsql/data
share_path=c:/dev/pgsql/share
PGPATH=C:/dev/pgsql/bin
ENCODING=
ENCODINGID=0
POSTGRES_SUPERUSERNAME=postgres
POSTGRES_BKI=c:/dev/pgsql/share/postgres.bki
POSTGRES_DESCR=c:/dev/pgsql/share/postgres.description
POSTGRESQL_CONF_SAMPLE=c:/dev/pgsql/share/postgresql.conf.sample
PG_HBA_SAMPLE=c:/dev/pgsql/share/pg_hba.conf.sample
PG_IDENT_SAMPLE=c:/dev/pgsql/share/pg_ident.conf.sample
The files belonging to this database system will be owned by user "postgres".
This user must also own the server process.

The database cluster will be initialized with locale English_United States.1252.

fixing permissions on existing directory C:/dev/pgsql/data ... ok
creating directory C:/dev/pgsql/data/global ... ok
creating directory C:/dev/pgsql/data/pg_xlog ... ok
creating directory C:/dev/pgsql/data/pg_clog ... ok
creating directory C:/dev/pgsql/data/pg_subtrans ... ok
creating directory C:/dev/pgsql/data/base ... ok
creating directory C:/dev/pgsql/data/base/1 ... ok
creating directory C:/dev/pgsql/data/pg_tblspc ... ok
selecting default max_connections ... 10
selecting default shared_buffers ... 50
creating configuration files ... ok
creating template1 database in C:/dev/pgsql/data/base/1 ... LOG:  could not open
 directory "/share/timezone": No such file or directory
child process was terminated by signal 5
initdb: failed
initdb: removing contents of data directory "C:/dev/pgsql/data"
Deleted file - C:\dev\pgsql\data\pg_hba.conf
Deleted file - C:\dev\pgsql\data\pg_ident.conf
Deleted file - C:\dev\pgsql\data\PG_VERSION
Deleted file - C:\dev\pgsql\data\postgresql.conf



Re: initdb crashing, signal 5, etc

From
Bruce Momjian
Date:
I saw the exact same behavior in my testing.

---------------------------------------------------------------------------

Jacob Hanson wrote:
> Had success with a snapshot from last week. Today's snapshot crashes
> several times during initdb. As soon as "selecting default
> max_connections" appears, postgres.exe crashes ("postgres.exe has
> encountered a problem and needs to close."). After closing 6 of those
> "10" appears and it continues to "selecting default shared_buffers".
> 11 crashes then "50" appears. It continues to the "No such file or
> directory" error and postgres.exe crashes once, then the signal 5
> message appears. Full initdb output is below.
>
> Observations: During the crashes, postgresql.conf is created, but
> remains 0 bytes. Cacls reports my "postgres" user has full access to
> the entire c:\dev\pgsql tree, including the data dir, postgresql.conf
> (after it has been created), and c:\dev\pgsql\share\timezone, which
> certainly exists (note: NOT "/shared/timezone" necessarily), etc.
>
>
>
>
>
>
> C:\dev\pgsql\data>initdb -d -L "c:/dev/pgsql/share"
> Running in debug mode.
> VERSION=7.5devel
> PGDATA=C:/dev/pgsql/data
> share_path=c:/dev/pgsql/share
> PGPATH=C:/dev/pgsql/bin
> ENCODING=
> ENCODINGID=0
> POSTGRES_SUPERUSERNAME=postgres
> POSTGRES_BKI=c:/dev/pgsql/share/postgres.bki
> POSTGRES_DESCR=c:/dev/pgsql/share/postgres.description
> POSTGRESQL_CONF_SAMPLE=c:/dev/pgsql/share/postgresql.conf.sample
> PG_HBA_SAMPLE=c:/dev/pgsql/share/pg_hba.conf.sample
> PG_IDENT_SAMPLE=c:/dev/pgsql/share/pg_ident.conf.sample
> The files belonging to this database system will be owned by user "postgres".
> This user must also own the server process.
>
> The database cluster will be initialized with locale English_United States.1252.
>
> fixing permissions on existing directory C:/dev/pgsql/data ... ok
> creating directory C:/dev/pgsql/data/global ... ok
> creating directory C:/dev/pgsql/data/pg_xlog ... ok
> creating directory C:/dev/pgsql/data/pg_clog ... ok
> creating directory C:/dev/pgsql/data/pg_subtrans ... ok
> creating directory C:/dev/pgsql/data/base ... ok
> creating directory C:/dev/pgsql/data/base/1 ... ok
> creating directory C:/dev/pgsql/data/pg_tblspc ... ok
> selecting default max_connections ... 10
> selecting default shared_buffers ... 50
> creating configuration files ... ok
> creating template1 database in C:/dev/pgsql/data/base/1 ... LOG:  could not open
>  directory "/share/timezone": No such file or directory
> child process was terminated by signal 5
> initdb: failed
> initdb: removing contents of data directory "C:/dev/pgsql/data"
> Deleted file - C:\dev\pgsql\data\pg_hba.conf
> Deleted file - C:\dev\pgsql\data\pg_ident.conf
> Deleted file - C:\dev\pgsql\data\PG_VERSION
> Deleted file - C:\dev\pgsql\data\postgresql.conf
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster
>

--
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

Re: initdb crashing, signal 5, etc

From
"Magnus Hagander"
Date:
I am specifically *not* seeing this on any of my three installes (1
normal + 2 vmware). I've tried with the exact same binary Merlin is
using (and he sees it), and he gets the crash and not me...

Further investigation definitly needed. Let's try to nail down where the
issue started showing up by pulling older cvs versions to identify the
commit. I wonder if this could be related to the pipe patch, since it
appears to show up on win32 only.

//Magnus

> -----Original Message-----
> From: Bruce Momjian [mailto:pgman@candle.pha.pa.us]
> Sent: Wednesday, July 28, 2004 9:48 AM
> To: Jacob Hanson
> Cc: pgsql-hackers-win32@postgresql.org
> Subject: Re: [pgsql-hackers-win32] initdb crashing, signal 5, etc
>
>
> I saw the exact same behavior in my testing.
>
> --------------------------------------------------------------
> -------------
>
> Jacob Hanson wrote:
> > Had success with a snapshot from last week. Today's
> snapshot crashes
> > several times during initdb. As soon as "selecting default
> > max_connections" appears, postgres.exe crashes ("postgres.exe has
> > encountered a problem and needs to close."). After closing
> 6 of those
> > "10" appears and it continues to "selecting default shared_buffers".
> > 11 crashes then "50" appears. It continues to the "No such file or
> > directory" error and postgres.exe crashes once, then the signal 5
> > message appears. Full initdb output is below.
> >
> > Observations: During the crashes, postgresql.conf is created, but
> > remains 0 bytes. Cacls reports my "postgres" user has full
> access to
> > the entire c:\dev\pgsql tree, including the data dir,
> postgresql.conf
> > (after it has been created), and c:\dev\pgsql\share\timezone, which
> > certainly exists (note: NOT "/shared/timezone" necessarily), etc.
> >
> >
> >
> >
> >
> >
> > C:\dev\pgsql\data>initdb -d -L "c:/dev/pgsql/share"
> > Running in debug mode.
> > VERSION=7.5devel
> > PGDATA=C:/dev/pgsql/data
> > share_path=c:/dev/pgsql/share
> > PGPATH=C:/dev/pgsql/bin
> > ENCODING=
> > ENCODINGID=0
> > POSTGRES_SUPERUSERNAME=postgres
> > POSTGRES_BKI=c:/dev/pgsql/share/postgres.bki
> > POSTGRES_DESCR=c:/dev/pgsql/share/postgres.description
> > POSTGRESQL_CONF_SAMPLE=c:/dev/pgsql/share/postgresql.conf.sample
> > PG_HBA_SAMPLE=c:/dev/pgsql/share/pg_hba.conf.sample
> > PG_IDENT_SAMPLE=c:/dev/pgsql/share/pg_ident.conf.sample
> > The files belonging to this database system will be owned
> by user "postgres".
> > This user must also own the server process.
> >
> > The database cluster will be initialized with locale
> English_United States.1252.
> >
> > fixing permissions on existing directory C:/dev/pgsql/data ... ok
> > creating directory C:/dev/pgsql/data/global ... ok creating
> directory
> > C:/dev/pgsql/data/pg_xlog ... ok creating directory
> > C:/dev/pgsql/data/pg_clog ... ok creating directory
> > C:/dev/pgsql/data/pg_subtrans ... ok creating directory
> > C:/dev/pgsql/data/base ... ok creating directory
> > C:/dev/pgsql/data/base/1 ... ok creating directory
> > C:/dev/pgsql/data/pg_tblspc ... ok selecting default
> max_connections
> > ... 10 selecting default shared_buffers ... 50 creating
> configuration
> > files ... ok creating template1 database in
> C:/dev/pgsql/data/base/1
> > ... LOG:  could not open  directory "/share/timezone": No
> such file or
> > directory child process was terminated by signal 5
> > initdb: failed
> > initdb: removing contents of data directory "C:/dev/pgsql/data"
> > Deleted file - C:\dev\pgsql\data\pg_hba.conf Deleted file -
> > C:\dev\pgsql\data\pg_ident.conf Deleted file -
> > C:\dev\pgsql\data\PG_VERSION Deleted file -
> > C:\dev\pgsql\data\postgresql.conf
> >
> >
> >
> > ---------------------------(end of
> > broadcast)---------------------------
> > TIP 4: Don't 'kill -9' the postmaster
> >
>
> --
>   Bruce Momjian                        |  http://candle.pha.pa.us
>   pgman@candle.pha.pa.us               |  (610) 359-1001
>   +  If your life is a hard drive,     |  13 Roberts Road
>   +  Christ can be your backup.        |  Newtown Square,
> Pennsylvania 19073
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to
> majordomo@postgresql.org
>
>

Older snapshots

From
Jacob Hanson
Date:
Are there older binary snapshots available? Kind of like pgAdmin3, how
there's the past few nights builds always available?

Jacob