Postgres DB crasht automatisch mit exit code 9

Hallo,

Ich nutze postgres als Datenbank Server unter Windows 2008 R2, seit der Installation kommt es immer wieder zu crashs mit exit code 9. Google spuckte leider nicht die gewünschten ergebnisse aus.
Vllt könnt ihr mir helfen. Es läuft ein Symantec Viren Scanner, wo aber die Postgress ordner von dem Scannen rausgenommen sind.

013-01-31 14:42:08 CETCONTEXT:  automatic vacuum of table "cloud.public.m201301_devicestatus"2013-01-31 14:43:14 CETLOG:  server process (PID 5820) exited with exit code 9
2013-01-31 14:43:14 CETLOG:  terminating any other active server processes
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:14 CETWARNING:  terminating connection because of crash of another server process
2013-01-31 14:43:14 CETDETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2013-01-31 14:43:14 CETHINT:  In a moment you should be able to reconnect to the database and repeat your command.
2013-01-31 14:43:15 CETLOG:  all server processes terminated; reinitializing
2013-01-31 14:43:24 CETFATAL:  pre-existing shared memory block is still in use
2013-01-31 14:43:24 CETHINT:  Check if there are any old server processes still running, and terminate them.

EDIT: es ist eine 32Bit installation von Postgres, alle anderen installation machen keine Probleme nur die.

Viele Grüße TK

Welche Postgres Version ist denn das?

Ausserdem würde ich den Viren Scanner testhalber mal komplett abschalten (oder sogar de-installieren).
Ein Virenscanner auf einem DB-Server ist eigentlich immer eine ganz schlechte Idee.

Das geht leider aus Firmen Politischen gründen nicht.
Es ist eine 9.1.7 32 Bit

Nicht mal zum Testen? Wie wollt ihr sonst den Viren Scanner als Bösewicht ausschliessen?
Man könnte den Server ja während der Zeit vom Netzwerk trennen wenn da jemand Bedenken hat.

Da es eine VM ist wird das Trennen vom netzwerk schwierig.
Was bedeutet denn exit code 9??

Da es eine VM ist wird das Trennen vom netzwerk schwierig.

Aber da ist doch das temporäre De-Installieren des Virenscanners noch einfacher.

  • Snapshot der VM erzeugen
  • Viren-Scanner de-installieren
  • Postgres starten und testen
  • VM wieder auf den Snapshot zurücksetzen.

Ich habe hier noch einen Hinweis auf eine möglicherweise falsche Einstellung für “shared_buffers” gefunden:
[PostgreSQL] PostgreSQL 9.0.4 crashing - Grokbase

Wenn das so einfach wäre, die VM_templates sind standart mäßig mit wem VirenScanner ausgestattet so ist eine installation ohne erst garnicht möglich.
De-Installieren kann ich denn nicht, dazu habe ich nicht die nötigen Rechte.

Das mit dem shared_buffer schaue ich mir mal an

Was ist denn ein “standart”? Oder meintest Du standard? (SCNR)

ich meine standard…

Hat sonst keiner ne idee?? Ich hab mittlerweile schon auf Postgres 9.2 gepatcht aber kein erfolg… crasht immer noch mit exit code 9 und weis nicht warum…

Ausser “VirenScanner de-installieren”? Nein.

Denn kann ich nicht Deinstallieren, Wobei ich noch nichtmal denke das es daran liegt.

Das werden wir wohl nie herausfinden.
Aber auf gefühlte 90% dieser Probleme die auf der Postgres Mailing Liste auftauchen, werden dadurch behoben.

ja aber das kann doch nicht die Lösung sein und dann einfach ein “ungeschützes” System zu haben.

Auf einem Datenbankserver sollte sowieso keine Software installiert werden, die nicht überprüft wurde. Der DB Server hat in einer sauberen Umgebung auch keinen Zugang zum Internet und dort darf auch niemand “einfach so” Software installieren. Deswegen ist ein Virenscanner dort nicht wirklich notwendig.

Du mußt jetzt sehr tapfer sein, aber: doch!


Wozu benötigst Du auf einem dedizierten Datenbankserver, der außer vielleicht SSH für den Admin nur die DB nach außen via TCP/IP anbietet, gleich noch einmal einen Virenscanner?


Andreas

da die Datenbank auf Windows läuft :smiley: :smiley: und meine vorgesetzten kein Linux derzeit wollen…

Auf linux und ssh wäre die Geschichte ja einfach gegessen… es muss aber unbedingt kompliziert sein…

Dann tausche SSH gegen RDP aus, oder nutze eine SSH-Implementierung für Windows. So what?

Andreas

Ich glaube wir schweifen grade vom Thema ab, SSH und RDP sind nicht das problem sondern der exit code 9.
Weis einer was exit code 9 bedeutet?

Ja, zurück zum Thema. Schalt Deine Virenschleuder, ähm, Scanner ab. Exit Code 9 unter Kleinweich heißt “The storage control block address is invalid”. Am besten also nicht nur abschalten, sondern deinstallieren. Noch besser: alles platt machen und gleich OHNE Software aus gelben Schachteln installieren.

Andreas