From Newsgroup: comp.databases.theory

I am starting a separate thread about this:

1. Visit this page
__ https://help.sap.com/viewer/product/SAP_ASE/16.0.4.0/en-US?task=whats_new_task
2. Select [ Download PDFs ] at top right
3. Choose the manuals you want, and download them.

Thanks, got them already.

Feel free to ask me questions.

What page size (2/4/8/16 KB) and what type of workload (Mixed/OLTP)
do you configure?

What are the strictly technical reasons to prefer ASE over SQL Server?
Or SQL Server over ASE?

How about this historical assessment:

https://dbdb.io/db/adaptive-server-enterprise

Buggy release, mismanagementrCa Doesn't sound like a product to go after ;-)

Nicola

--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.databases.theory

On Wednesday, 4 August 2021 at 21:21:38 UTC+10, Nicola wrote:
I am starting a separate thread about this:

1. Visit this page
__ https://help.sap.com/viewer/product/SAP_ASE/16.0.4.0/en-US?task=whats_new_task
2. Select [ Download PDFs ] at top right
3. Choose the manuals you want, and download them.

Thanks, got them already.

Feel free to ask me questions.

Whoa. Evidently you have missed the context. You said you downloaded Sybase, and that you wanted to benchmark and confirm some declarations that I made, while re-framing it as "claims". I said I encourage any academic moving from their isolated tiled room, into the real world. I gave some cautions, I offered assistance, in that regard. That is to get you merrily on your way, on your stated task.
My invitation should be not be construed as "feel free to ask me questions about anything you want"
Further, you have to do some work and climb the learning curve yourself.

What page size (2/4/8/16 KB) and what type of workload (Mixed/OLTP)
do you configure?

You have found the manuals. You need to find the SAP Developers Forum or network. Questions of this type, which have many considerations, are discussed and answered there. There are a few SAP/Sybase engineers there, who answer questions.
SG was a Sybase Partner for 17 years, right up to the acquisition by SAP. I was active in the previous Sybase technical forum, and there are many posts (with full discussion) on this and related subjects. You could have just googled the subject plus my name, and you would get your answer. That was 20 years ago. Truth does not change, the answers are still valid.
After SAP, that forum has been removed from the internet, it is now archived, in the new SAP developer forum (you can no longer google for the subject matter). Please go there (use the search facility) and read the discussion for considerations, and choose from the answers. Ask away.
Sorry, I do not have the time to post duplicate info, or to get into yet another open discussion, about closed subjects.
On the other hand, if you have a specific question re your stated project, I would be happy to help. No discussion, just the answer.

What are the strictly technical reasons to prefer ASE over SQL Server?
Or SQL Server over ASE?

Ditto.

How about this historical assessment:

https://dbdb.io/db/adaptive-server-enterprise

These days, anyone with a keyboard; two fingers and a bit of tissue connecting those two fingers, can post articles on the internet. They do not have to have any grey cells in that connective tissue, or any actual experience about the subject. Then there is a host of academics, who post negative articles against anything they do not understand, which as you have seen in the last ten years, is an awful lot, due to their proud isolation from reality. Such people, both ordinary idiots, and academically qualified idiots, are unskilled, and worse, unaware that they are unskilled. They think themselves skilled.
There are a good few scientific papers on the subject (as distinct from the filth that academics in this field produce). This is required reading for anyone who takes my courses. It defines the state of academics in this field, as evidenced in our threads.
__ Unskilled and Unaware 1999, Kr|+ger & Dunning
____ https://www.softwaregems.com.au/Documents/Reference/Unskilled%20%26%20Unaware%201999.pdf
__ Unskilled and Unaware 2008, follow-up, responding to the attacks
____ https://www.softwaregems.com.au/Documents/Reference/Unskilled%20%26%20Unaware%202008.pdf
Hence the internet is a cesspool, and as a number of people have stated, it is rare to find a genuine authority on a subject, where answers are direct and permanent, instead of endless discussion without resolution.
Choose what you read carefully.

Buggy release, mismanagementrCa Doesn't sound like a product to go after ;-)

You are right. Drop it. Go back to your freeware herd, the hundreds of programs that pretend to be a server, that pervert SQL, that re-define concepts in order for their broken implementation to appear compliant. Given your recent questions, and your difficulty installing and running commercial products, you will be so much happier there. No need to obtain actual experience in the real world, just rail against it from the safety of the tiled cell.
Cheers
Derek
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.databases.theory

On Thursday, 5 August 2021 at 12:56:18 UTC+10, Derek Ignatius Asirvadem wrote:

On Wednesday, 4 August 2021 at 21:21:38 UTC+10, Nicola wrote:

What page size (2/4/8/16 KB) and what type of workload (Mixed/OLTP)
do you configure?

On the other hand, if you have a specific question re your stated project, I would be happy to help. No discussion, just the answer.

In case you wish to leave the asylum for an hour or so, for a picnic in the park, to enjoy the sunshine, and perhaps actually produce something on the shiny new SAP/Sybase Server ... if I rephrase your question such that it is pertinent and direct, such as:
< < What page size (2/4/8/16 KB) do you recommend that I configure ?
Because this is a permanent physical article, that precedes installation, for what I expect is best for you, eg. I expect you to stress the server with benchmarks and large Transactions, as well as a mixed OLTP+OLAP load = 4KB.
If you are on Unix/Linux, make sure you create only Raw Partitions for all Devices, never filesystem files.
Cheers
Derek
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.databases.theory

On 2021-08-06, Derek Ignatius Asirvadem <derek.asirvadem@gmail.com> wrote:

if I rephrase your question such that it is pertinent and direct, such
as:

< < What page size (2/4/8/16 KB) do you recommend that I configure ?

Because this is a permanent physical article, that precedes
installation, for what I expect is best for you, eg. I expect you to
stress the server with benchmarks and large Transactions, as well as
a mixed OLTP+OLAP load = 4KB.

Ok, thanks. I had inferred from your documents that your benchmarks were
done on systems configured with 2 KB pages, so I was wondering whether
(that is the case and) you had a point to always prefer smaller
pages to larger ones.

If you are on Unix/Linux, make sure you create only Raw Partitions for
all Devices, never filesystem files.

Ok.

Nicola
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.databases.theory

On Sunday, 8 August 2021 at 22:24:39 UTC+10, Nicola wrote:

On 2021-08-06, Derek Ignatius Asirvadem wrote:
if I rephrase your question such that it is pertinent and direct, such
as:

< < What page size (2/4/8/16 KB) do you recommend that I configure ?

Because this is a permanent physical article, that precedes
installation, for what I expect is best for you, eg. I expect you to stress the server with benchmarks and large Transactions, as well as
a mixed OLTP+OLAP load = 4KB.

Ok, thanks. I had inferred from your documents that your benchmarks were done on systems configured with 2 KB pages,

Yes.
TPC requires small fast Transactions.

so I was wondering whether
(that is the case and)

Yes.

you had a point to always prefer smaller
pages to larger ones.

It is not a "preference", it is a scientifically determined article. We have over 250 configuration parms, most of which are related to each other, additionally all resources [look into the named caches] are configured based on memory; etc. We use a spreadsheet and everything. Even for pure P&T assignments, I publish the spreadsheet as an appendix in the final Before & After report.
Yes.
- For myself.
- For servers & databases that I rebuild for customers AFTER writing a Version2 system for them, which means the entire app, and small fast Transactions in stored procs only.
For servers & databases that I rebuild but not rewrite, ie. no change to the app or the database logically, no. I leave the pagesize as is.
For correction of mistakes (eg. customer has moved from 2KB to 4KB to 8KB to ease Transaction problems, which is a stupid thing to do, as it has no effect on the problem because it is not the cause of the problem), as part of the server rebuild, I determine what it should be for their usage; load, vs their physical resources, and implement that. No, you can't have the spreadsheet.
Most Sybase customers are on 4KB.
High-end OLTP customers (serviced by high-end guys such as me) are on 2KB. Ie. 2KB has never been an issue.
Some are on 8KB for a mixed load.
16KB is for OLAP only.
Advice for you:

Because this is a permanent physical article, that precedes
installation, for what I expect is best for you, eg. I expect you to stress the server with benchmarks and large Transactions,

Guaranteed that you will jack around with huge Transactions.
Guaranteed you will think the problem is 2KB pagesize.
Guaranteed that you will not listen to reasons, as to why that is false. Argument avoided.

as well as
a mixed OLTP+OLAP load = 4KB.

Regardless of what you think, that will be the the best for the general load, over any duration of time.
Cheers
Derek
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.databases.theory

Derek,
I've gone full speed with some benchmark scripts up and running against
a pretty default ASE installation. I have created two devices, one for databases ("userdbdev") and one for the transaction log ("userlogdev").
My test database was created with:

create database scratch on userdbdev = '1g' log on userlogdev = '1g'

After several runs of my scripts, I have started to get this error:

The transaction log in database scratch is almost full. Your transaction
is being suspended until space is made available in the log.

1. How do I flush the transaction log?
2. How do I monitor the size of the transaction log?
3. How do I avoid the above message in the first place?

Nicola

--- Synchronet 3.21d-Linux NewsLink 1.2