Firstly I would like to second Dossy on the quality of the aolserver
community. In addition I would like to say that one of the biggest
reasons why the aolserver community has life in 2006 and is of such a
high calibre is because of Dossy himself.
I also want to say that I was joking in the
Mark Aufflick said:
I also want to say that I was joking in the comment quoted below. Weak
technology (especially expensive weak technology) gets me a little hot
under the collar sometimes but I don't think there was any real
In that regards, IBM to me has always been a shocking example with
Mat Kovach said:
I see many large scale sites that run UDB/DB2. I know of at least three
sites that have, easily, 1000 UDB instances running. I lot of people
What kind of sites are they? What industries?
Cheers,
Bas.
--
AOLserver - http://www.aolserver.com/
To Remove yourself from this
On Mon, 2006-04-10 at 13:47 +0100, Bas Scheffers wrote:
Who uses DB2? Is it any good, or much like IBM's other software?
Just about every IBM midrange system comes with DB2. There are an awful lot of AS/400 or iSeries out there and there were a lot of System/36 and System/38's as well.
Bas Scheffers wrote:
Mat Kovach said:
I see many large scale sites that run UDB/DB2. I know of at least three
sites that have, easily, 1000 UDB instances running. I lot of people
What kind of sites are they? What industries?
Large data centers, mainly for finance and logistics companies.
I've seen it in the wild in financial institutions (mainly banks)
running on as/400 or iSeries boxes. Most would then use products like
datamirror to migrate it out of db2 and into sql server, etc.. to run
their reporting packages against.
best,
Joe Kondel
On Apr 10, 2006, at 10:33 AM, Mat
On 2006.04.07, Mark Aufflick [EMAIL PROTECTED] wrote:
Although that's not a dissimilar problem to the NULL as empty string
issue in the Oracle C libraries (which really really bites).
Um, where are you getting this information from? Oracle's OCI libraries
definitely expose whether a column
Mark Aufflick said:
I feel a good old fashioned usenet style flame war coming on =)
Slashdot has flame wars, all we have are friendly argument with as outcome
learning more about our favourite and other products.
In Sybase, NULL matches NULL in a search clause (by default - as
Yes. I personaly
On 2006.04.07, Bas Scheffers [EMAIL PROTECTED] wrote:
Mark Aufflick said:
I feel a good old fashioned usenet style flame war coming on =)
Slashdot has flame wars, all we have are friendly argument with as outcome
learning more about our favourite and other products.
I actually wanted to
Most of the web world runs on MySQL and does ok - just like CDBaby who
You can't possibly liken Sybase to MySQL!
I know that they are not even in the same league technically, but they
do show disturbingly similar philosophies - like making NULL = NULL by
default because many clients with
That's great - I'm a bit of a Sybase newby and this tip makes life a lot easier.
Given that this is so easy and reasonable I don't understand why so
many Sybase developers employ other unreliable methods to emulate
triggers.
On 4/5/06, Bas Scheffers [EMAIL PROTECTED] wrote:
select @@identity
Mark Aufflick said:
do show disturbingly similar philosophies - like making NULL = NULL by
default because many clients with poorly trained developers asked for
I recon that's the only one! (and I still don't see why this is such a bad
thing) My problem with MySQL isn't this simplification, it's
On 2006.04.06, Mark Aufflick [EMAIL PROTECTED] wrote:
That's great - I'm a bit of a Sybase newby and this tip makes life a
lot easier.
Given that this is so easy and reasonable I don't understand why so
many Sybase developers employ other unreliable methods to emulate
triggers.
For the
Dossy Shiobara said:
You're right; MySQL is orders of magnitude better than Sybase. I'm dead
Yeah, when it doesn't blow up in weird and wonderul ways... And you don't
mind not being able to do online backups And if you think '-00-00'
or '2006-02-30' is a date... (oh no, just tested mysql
On 2006.04.06, Bas Scheffers [EMAIL PROTECTED] wrote:
Dossy Shiobara said:
You're right; MySQL is orders of magnitude better than Sybase. I'm dead
Yeah, when it doesn't blow up in weird and wonderul ways... [...]
I'll bet a nickel it's user error (whether that user is a developer or
the
It sounds like you guys are comparing rotten oranges and rotten apples.
tom jackson
On Thursday 06 April 2006 05:37, Bas Scheffers wrote:
Dossy Shiobara said:
You're right; MySQL is orders of magnitude better than Sybase. I'm dead
Yeah, when it doesn't blow up in weird and wonderul ways...
Tom Jackson said:
It sounds like you guys are comparing rotten oranges and rotten apples.
Are there any apples that aren't considered rotten by someone, somewhere?
Like I said before, every RDBMS has it's issues...
...MySQL just takes them to a whole new level and gets even otherwise
sensible
Dossy Shiobara said:
Oh man, who is feeding you this pack of lies? I mean, MySQL has been
able to do hot online backups since May 2002! Well, you've been able to
How? All I can find is this mysqldump and mysqlhotcopy, which lock
tables while they are being dumped. Hardly online backup if you
On 2006.04.06, Bas Scheffers [EMAIL PROTECTED] wrote:
The difference? If I *had* to, I could extend MySQL to do exactly what
I need. (Beware: Tcl as a supported UDF language for MySQL stored
procs! Muwahaha.)
Why waste your time, Postgres already has that! :) (and loads of other
On 6 Apr 2006, at 17:42, Dossy Shiobara wrote:
Can you embed Postgres? Before I got turned on to SQLite, MySQL
served
I don't see that as a downside, it's not what I use it for. Like you
say, there's SQLite for that. Try embedding MySQL into a non-GPL
project and see how free it is then!
On 2006.04.06, Bas Scheffers [EMAIL PROTECTED] wrote:
As much as people criticize MySQL's replication and clustering, where's
Postgres's? I don't see any mention of it in Postgres 8.1 docs:
http://www.postgresql.org/docs/8.1/interactive/index.html
That's because replication isn't part
On 6 Apr 2006, at 19:13, dhogaza@PACIFIER.COM wrote:
The people who wrote MySQL have the ethical standards of the George W.
Bush administration.
What, it's made by the same folks as those behind JBoss!? ;-)
Bas.
--
AOLserver - http://www.aolserver.com/
To Remove yourself from this list,
On 4/6/06, Dossy Shiobara [EMAIL PROTECTED] wrote:
Did you know that Sybase silently promotes an empty string to a string
of 1 character? You can't actually store an empty string in a Sybase
database. That is absolutely ridiculous. I'd expect that kind of
behavior from some college-level
On 4/6/06, Bas Scheffers [EMAIL PROTECTED] wrote:
Mark Aufflick said:
do show disturbingly similar philosophies - like making NULL = NULL by
default because many clients with poorly trained developers asked for
I recon that's the only one! (and I still don't see why this is such a bad
On 4/7/06, Mark Aufflick [EMAIL PROTECTED] wrote:
On 4/6/06, Dossy Shiobara [EMAIL PROTECTED] wrote:
Did you know that Sybase silently promotes an empty string to a string
of 1 character? You can't actually store an empty string in a Sybase
database. That is absolutely ridiculous. I'd
select @@identity for the last identity value assigned in the
current session. This is acros all tables with id columns, so if you
are inserting multiple in one transaction and need all of them, be
sure to assign them to another variable after each insert.
Cheers,
Bas.
On 5 Apr 2006, at
Hi Mark,
* Mark Aufflick [EMAIL PROTECTED] [060405 06:31]:
Sybase, like SQLlite, does (now) support an auto-incrementing column
type, but AFAICT there is no way to find out which id you just
inserted.
Sybase provides the @@identity variable which holds the last
Mark Aufflick said:
I have seen some large Sybase databases go down quite spectacularly.
Any details on how or why? One common cause for downtime is filling up the
log space and locking up the DB, not knowing how to fix it. But they
should have read the manual!
Most of the web world runs on
Can't you write a function to do the insert and return the value of the oid?
If you are contemplating adding a new query, seems like it would be just as
easy to replace the insert statement with a function call.
Another possibility is to grab a number of oids and cache them in your
Mark Aufflick said:
Doesn't that still serialize all updates requiring access to that
sequence?
Sequences are used for inserts, not updates. Well, I haven't seen them
being used in updates anyway.
If you used it within a transaction it would also have implications
with other transactions
On Tue, Apr 04, 2006 at 03:02:21PM +1000, Mark Aufflick wrote:
Doesn't that still serialize all updates requiring access to that sequence?
No. Sequences in an RDBMS are designed to scale gracefully under
heavy concurrent load. (This is basic stuff, grab any good
intro. book on databases and
On 2006.04.04, Bas Scheffers [EMAIL PROTECTED] wrote:
All Systems have their merrits. As (commercial) enterprise database go, I
have _much_ better experiences with Sybase than Oracle for ease of
maintainance (best backups in the business), optimization (it just uses
indexes that make sense
Andrew Piskorski said:
No. Sequences in an RDBMS are designed to scale gracefully under
heavy concurrent load. (This is basic stuff, grab any good
I think Mark was talking about my Sybase solution (sequence table) for if
you don't want to use identity columns.
I think SQL Server and Sybase
Dossy Shiobara said:
I just have to ask: what version of Sybase and Oracle are you comparing?
9i and 11.92, 12, 12.5.
The worst Oracle instance was maintained by a couple of very good DBAs, at
least according to the guy who actually worked for Oracle and did an audit
of our system. They were
Hi Andrew,
I might have missed the context in my reply - we were discussing
Sybase which is one of those databases missing sequence support.
We are a little off topic, but Bas was suggesting a reasonable way
around the missing sequences (very reasonable compared with many other
attempts to work
I've added so many 2c this thread is nearly tax deductable!
.. I have never actually seen a Sybase instance go down
or mess up my data.
I have seen some large Sybase databases go down quite spectacularly.
The whole financial world seems to run on Sybase, and so far it's been
doing OK...
Doesn't that still serialize all updates requiring access to that sequence?
If you used it within a transaction it would also have implications
with other transactions using that sequence especially in a rollback
situation.
In the context of AOLServer you could minimise the impact of these
William Scott Jordan said:
ns_db dml $db INSERT INTO test (test_column) SELECT
COALESCE(MAX(test_column),0) + 1 FROM test
Ehrm, that is a very, very bad way of creating keys. It is no unlikely
that on a very loaded site two people might insert at the same time and
get the same key!
The correct
The way we do this in OpenACS is to have a primary key in each table
that is populated from a sequence. So in your code you get the next
number from the sequence, do the insert (storing the number in the
primary key column), and then you have that number already in your
possession to use
39 matches
Mail list logo
