Package: postgresql-common Version: 100 Severity: normal After an apparently successfull conversion the db wouldn't start:
-------- Copying old configuration files... Copying old start.conf... Copying old pg_ctl.conf... Stopping target cluster... Stopping old cluster... Disabling automatic startup of old cluster... Configuring old cluster to use a different port (5434)... Starting target cluster on the original port... The PostgreSQL server failed to start. Please check the log output. Error: Could not start target cluster; please check configuration and log files -------- Postgres didn't start so there are no postgres logs to check. While checking through the config files I found the following in the converted postgresql.conf: external_pid_file = '/var/run/postgresql/8.4-mydb.pid' # write an extra PID file external_pid_file = '/var/run/postgresql/8.3-mydb.pid' # write an extra PID file (I have no "main" cluster by default, although one was (successfully) created (and started) as part of the new install). That's very minor though, and wasn't the cause of the problem. What was was found in syslog and was caused by a change in the format of pg_hba.conf: "authentication option not in name=value format" I'm using ident maps for local authentication. Previously the authentication options were simply a pair with no = sign required. It would be nice if the conversion could check for this (anything following a known auth method on a line has to be an auth option, which is now required to be in name=value format), and warn the user if it finds any auth options that don't include an = sign. (You could probably just insert an = sign between the 2 words and handle the conversion automatically, but the user should still be warned in that situation). Failing that it would be nice to include this information in a news item (although that would probably go with the newer postgres version rather than postgresql-common). Once I corrected that everything started up just fine. I suspect that this isn't likely to hit most people, and is easy to fix once you track it down - but isn't as seamless as it could be. And if you happen to not look at syslog can be tricky to track down. Would be nice if the postmaster output was passed through to the console in case of a startup failure. Thanks in advance, -jcm -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (900, 'testing'), (100, 'unstable'), (1, 'experimental') Architecture: amd64 (x86_64) Kernel: Linux 2.6.30.4+00.amd64.fbcondecor (SMP w/2 CPU cores; PREEMPT) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages postgresql-common depends on: ii adduser 3.110 add and remove users and groups ii debconf [debconf-2.0] 1.5.27 Debian configuration management sy ii lsb-base 3.2-22 Linux Standard Base 3.2 init scrip ii lsb-release 3.2-22 Linux Standard Base version report ii postgresql-client-common 100 manager for multiple PostgreSQL cl ii procps 1:3.2.8-1 /proc file system utilities Versions of packages postgresql-common recommends: ii ssl-cert 1.0.23 simple debconf wrapper for OpenSSL postgresql-common suggests no packages. -- debconf information: * postgresql-common/obsolete-major: -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
