Powergres (was Re: Errors on restoring a 7.1.2 dump to 7.3.3)

On restoring a plain-text SQL script file dumped from a version 7.1.2
database to 7.3.3 I got the following errors.

ERROR: copy: line 67568, Temporary log file is full
lost synchronization with server, resetting connection
ERROR: copy: line 66379, Temporary log file is full
lost synchronization with server, resetting connection
ERROR: copy: line 54948, Temporary log file is full
lost synchronization with server, resetting connection

..and many

ERROR: Default table space is full

...errors.

Could any one tell me what this is indicative of and perhaps how to fix
it?

Many thanks as ever.

Stuart Woodward <woodward(at)garage(dot)co(dot)jp> writes:
> On restoring a plain-text SQL script file dumped from a version 7.1.2
> database to 7.3.3 I got the following errors.

> ERROR: copy: line 67568, Temporary log file is full
> ..and many
> ERROR: Default table space is full

You're joking, no? Neither string appears anywhere in the 7.3 sources,
nor any other PG release that I know of.

regards, tom lane

On Wed, 03 Sep 2003 23:47:37 -0400
Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> You're joking, no? Neither string appears anywhere in the 7.3 sources,
> nor any other PG release that I know of.4

Apologies. I rechecked with my minion and the version being used is a
variant called "Powergres" by Fujitsu.

(Japanese page)
http://software.fujitsu.com/jp/symfoware/postgresqlplus/index.html

I'll take it up with them. Thanks for checking that out though.

On Thu, 2003-09-04 at 00:17, Stuart Woodward wrote:
> On Wed, 03 Sep 2003 23:47:37 -0400
> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
[snip]
> Apologies. I rechecked with my minion and the version being used is a
> variant called "Powergres" by Fujitsu.
>
> (Japanese page)
> http://software.fujitsu.com/jp/symfoware/postgresqlplus/index.html

What makes Powergres special? From looking at the pretty pictures
(and ignoring the squiggly lines), it looks like replication.

--
-----------------------------------------------------------------
Ron Johnson, Jr. ron(dot)l(dot)johnson(at)cox(dot)net
Jefferson, LA USA

"they love our milk and honey, but preach about another way of living"
Merle Haggard, "The Fighting Side Of Me"

On Wednesday, September 3, 2003, at 10:47 PM, Tom Lane wrote:

> Stuart Woodward <woodward(at)garage(dot)co(dot)jp> writes:
>> On restoring a plain-text SQL script file dumped from a version 7.1.2
>> database to 7.3.3 I got the following errors.
>
>> ERROR: copy: line 67568, Temporary log file is full
>> ..and many
>> ERROR: Default table space is full

We recently ran into a nasty Oracle bug at work where the default
tablespace was full. It just died, and was unrecoverable without
restoring from backup.

> You're joking, no? Neither string appears anywhere in the 7.3 sources,
> nor any other PG release that I know of.
>
> regards, tom lane
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 9: the planner will ignore your desire to choose an index scan if
> your
> joining column's datatypes do not match

Ron Johnson <ron(dot)l(dot)johnson(at)cox(dot)net> writes:
> What makes Powergres special? From looking at the pretty pictures
> (and ignoring the squiggly lines), it looks like replication.

Not sure if there's any replication in it. My understanding is that
it amounts to ripping out the lowest layers of the Postgres code and
bolting in Fujitsu's proprietary storage engine (originally developed
for some other DBMS) instead. The Fujitsu people say that this gives
some significant performance improvements, and for all I know it might
do replication too. But I've not yet seen a straight answer about
whether their code preserves MVCC semantics, rollback-able DDL changes,
and other nice stuff that the regular PG code provides. This report
from Stuart isn't exactly giving me warm fuzzies about the robustness
of Fujitsu's code, either.

regards, tom lane

On Thu, 2003-09-04 at 01:57, Tom Lane wrote:
> Ron Johnson <ron(dot)l(dot)johnson(at)cox(dot)net> writes:
> > What makes Powergres special? From looking at the pretty pictures
> > (and ignoring the squiggly lines), it looks like replication.
>
> Not sure if there's any replication in it. My understanding is that
> it amounts to ripping out the lowest layers of the Postgres code and
> bolting in Fujitsu's proprietary storage engine (originally developed
> for some other DBMS) instead. The Fujitsu people say that this gives
> some significant performance improvements, and for all I know it might
> do replication too. But I've not yet seen a straight answer about
> whether their code preserves MVCC semantics, rollback-able DDL changes,
> and other nice stuff that the regular PG code provides. This report
> from Stuart isn't exactly giving me warm fuzzies about the robustness
> of Fujitsu's code, either.

Here's what someone sent me off-list:
http://osb.sra.co.jp/PowerGres/function-en.php

--
-----------------------------------------------------------------
Ron Johnson, Jr. ron(dot)l(dot)johnson(at)cox(dot)net
Jefferson, LA USA

"Man, I'm pretty. Hoo Hah!"
Johnny Bravo