Обсуждение: Bug #783: WINDOWS 2000: invalid primary checkpoint record

Поиск
Список
Период
Сортировка

Bug #783: WINDOWS 2000: invalid primary checkpoint record

От
pgsql-bugs@postgresql.org
Дата:
Le Quang Anh Ngoc (ngoc@unico-t.co.jp) reports a bug with a severity of 2
The lower the number the more severe it is.

Short Description
WINDOWS 2000: invalid primary checkpoint record

Long Description
Hello,
I'm using Windows 2000, PostgreSQL 7.2.1-win32.

While System was be running, the power was interrupted and rebooted. Then I run pgsvmgr.exe again, from client machine
Itry to run psql.exe and login but I cannot.  

Please, help me to restore the data.
Thanks alot.



Sample Code
*The content of file pg.errors :

DEBUG:  database system was interrupted at 2002-09-20 21:03:14 JST
DEBUG:  ReadRecord: out-of-sequence SUI 2 (after 3) in log file 1, segment 108, offset 8863744
DEBUG:  invalid primary checkpoint record
DEBUG:  open of C:\postgresql-7.2.1-win32\bin\..\data\pg_xlog/0000000100000021 (log file 1, segment 33) failed: No such
fileor directory 
DEBUG:  invalid secondary checkpoint record
FATAL 2:  unable to locate a valid checkpoint record


No file was uploaded with this report

FATAL 1: LWLockAcquire: can't wait without a PROC structure

От
fredrik chabot
Дата:
Hi,

I get the following msg "FATAL 1:  LWLockAcquire: can't wait without a
PROC structure" in the log file about once a day and every user is
kicked out of the database.

Running on a single cpu amd with Redhat Linux 7.2 kernel 2.4.7-10 and
the postgresql-7.2.1-5 rpm's

NB I tried running it on a dual cpu smp machine running 2.2.18* and
there I got this msg far more frequently about once an hour.

usaly 30-50 connections during the daytime.

:-) help!

regards

fredrik chabot

Re: FATAL 1: LWLockAcquire: can't wait without a PROC structure

От
Tom Lane
Дата:
fredrik chabot <fredrik@f6.nl> writes:
> I get the following msg "FATAL 1:  LWLockAcquire: can't wait without a
> PROC structure" in the log file about once a day and every user is
> kicked out of the database.

We found the cause of that just a couple days ago.  It'll be fixed in
7.3beta3.

            regards, tom lane