Re: Problem with JDBC interface in Postgresql-6.3
[ I'm forwarding this to the hackers list, as this looks like a problem in
the backend ]
On Mon, 16 Mar 1998, Uri Blumenthal wrote:
I've tried to reproduce this, but it's not failing.
Lucky you are...
This sounds like the backend dying while in the middle of the fastpath
call. If you run the backend with debugging enabled, does anything appear
when this fault occurs?You bet. Here's the output of "-d 3 -o /tmp/pgsql.log":
NOTICE: DateStyle is Postgres with US (NonEuropean) conventions
NOTICE: buffer leak [77] detected in BufferPoolCheckLeak()
NOTICE: buffer leak [77] detected in BufferPoolCheckLeak()
NOTICE: buffer leak [77] detected in BufferPoolCheckLeak()
NOTICE: buffer leak [77] detected in BufferPoolCheckLeak()
Failed Assertion("!(BufferIsValid(buffer)):", File: "bufmgr.c", Line: 1294)
!(BufferIsValid(buffer)) (0) [No such file or directory]
Uri, what platform are you running the backend? This looks like something
serious is going wrong here...
And here's the Java dump:
[ dump snipped - it's just confirming the above ]
--
Peter T Mount petermount@earthling.net or pmount@maidast.demon.co.uk
Main Homepage: http://www.demon.co.uk/finder
Work Homepage: http://www.maidstone.gov.uk Work EMail: peter@maidstone.gov.uk
Import Notes
Reply to msg id not found: 199803162353.SAA01142@katana.watson.ibm.com
On Mon, 16 Mar 1998, Uri Blumenthal wrote:
"Peter" == Peter T Mount <pmount@maidast.demon.co.uk> writes:
Peter> I've tried to reproduce this, but it's not failing.
Darn... I recompiled and re-installed PostgreSQL-6.3, and
now I can't reproduce this yesterday's error either. (:-([but see below.]
Opening large object 19201
Opening test destination object
Copying large object to file
object size=6839
copied 512 bytes, 6327 left
copied 1024 bytes, 5815 left
copied 1536 bytes, 5303 left
copied 2048 bytes, 4791 left
copied 2560 bytes, 4279 left
copied 3072 bytes, 3767 left
copied 3584 bytes, 3255 left
copied 4096 bytes, 2743 left
copied 4608 bytes, 2231 left
copied 5120 bytes, 1719 left
copied 5632 bytes, 1207 left
copied 6144 bytes, 695 left
copied 6656 bytes, 183 left
copied 6839 bytes, 0 left
The whole object copied successfully (6839 bytes).
Closing object
weird... at least that problems now gone ;-)
But now I'm trying "example.datestyle" and it fails me:
PostgreSQL datestyle test v6.3 rev 1
Connecting to Database URL = jdbc:postgresql:foo
Connected...Now creating a statementRunning tests:
Test 0 - postgres,european failed, returned 08-01-1998
Test 1 - postgres,us failed, returned 01-08-1998
Test 2 - iso failed, returned 1998-01-08
Test 3 - sql,european failed, returned 08/01/1998
Test 4 - sql,us failed, returned 01/08/1998
Test 5 - german failed, returned 08.01.1998
Now closing the connection
Just ran here:
Running tests:
Test 0 - postgres,european passed, returned 08-01-1998
Test 1 - postgres,us passed, returned 01-08-1998
Test 2 - iso passed, returned 1998-01-08
Test 3 - sql,european passed, returned 08/01/1998
Test 4 - sql,us passed, returned 01/08/1998
Test 5 - german passed, returned 08.01.1998
Now closing the connection
Surprising thing is, that the date appears to be the same,
so it's beyond me why
standard.equals(rs.getDate(1))returns "False".
Any idea?
Again, are your client & backend on different machines/platforms, and what
platform(s) are you using?
--
Peter T Mount petermount@earthling.net or pmount@maidast.demon.co.uk
Main Homepage: http://www.demon.co.uk/finder
Work Homepage: http://www.maidstone.gov.uk Work EMail: peter@maidstone.gov.uk
Import Notes
Reply to msg id not found: 199803170332.WAA00766@alisan.ibm.net | Resolved by subject fallback