Re: pg_upgrade v9.1 issues

lalong wrote:
> I have 2 instances of postgres installed on same machine under Vista and have
> run into the following issues when trying to migrate my data from one
> instance into the other....
>
> 1) If the upgrade process fails, the postmaster.tid file may remain even
> though the postmaster service is no longer running. I have to manually
> delete it to re-try. This may occur in either the source or target area.

That is odd --- even if the pid file remains, it shouldn't prevent the
new server from starting. Can you reproduce it using manual commands or
is it something specific to pg_upgrade? -l log will show you the
commands pg_upgrade is running.

> 2) Recommend adding additional parameters to allow specifying target and
> source roles & pw's. I was getting errors when the same role does not exist
> in both instances.

Well, I am not sure how that happened if the new cluster was empty on
start. Can you be more specific?

> 3) Received the following error
> *********
> C:\Program Files\PostgreSQL\9.1\bin>pg_upgrade --old-datadir
> "D:\servoy6\application_server\database" --new-datadir "c:\program
> files\postgresql\9.1\data" --old-bindir
> "D:\servoy6\application_server\postgres_db\bin" --new-bindir "c:\program
> files\postgresql\9.1\bin"
> Performing Consistency Checks
> -----------------------------
> Checking current, bin, and data directories ok
> Checking cluster versions ok
> Checking database user is a superuser ok
> Checking for prepared transactions ok
> Checking for reg* system oid user data types ok
> Checking for contrib/isn with bigint-passing mismatch ok
> Creating catalog dump ok
>
> There were problems executing
> ""D:\servoy6\application_server\postgres_db\bin/pg_ctl" -w -l "nul" -D
> "D:\servoy6\application_server\database" stop >> "nul" 2>&1"
> Failure, exiting
> *********

That is odd. Does it work if you run it manually?

> Note: I don't know if this matters, but the target (new v9.1) source has
> instance set to start automatically in services. The source (old v9.0) is
> not. The error is referring to the source instance. Apparently, it is
> trying to shut the service down and if it is already down, then it errors
> out as opposed to continuing on.

Well, both servers should be down on start, as mentioned in the
pg_upgrade documentation. I don't see how the auto-start would affect
that.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +