Sybase NNTP forums - End Of Life (EOL)

The NNTP forums from Sybase - forums.sybase.com - are now closed.

All new questions should be directed to the appropriate forum at the SAP Community Network (SCN).

Individual products have links to the respective forums on SCN, or you can go to SCN and search for your product in the search box (upper right corner) to find your specific developer center.

high-performance wish list: no logging option for a device

9 posts in Product Futures Discussion Last posting was on 2002-03-28 15:39:48.0Z
Pablo Sanchez Posted on 2002-01-12 22:34:28.0Z
From: "Pablo Sanchez" <pablo@dev.null>
Subject: high-performance wish list: no logging option for a device
Date: Sat, 12 Jan 2002 15:34:28 -0700
Lines: 18
Organization: High-Performance Database Engineering
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <MvxQ#n7mBHA.297@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 207.225.105.222
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:942
Article PK: 95182

Howdy!

I'd like to suggest that Sybase Engineering offer the ability to
create and alter devices so logging is off or on per device. This
would allow me set aside certain devices with no logging set, and
assign them to say, tempdb.

It might be a misnomer to say 'no logging' in that we might wish to
perform some minimal logging.

Thx!
--
Pablo Sanchez, High-Performance Database Engineering
www.hpdbe.com
Independent Contractor, available for short-term and long-term
contracts


Bret Halford Posted on 2002-01-13 17:26:09.0Z
Message-ID: <3C41C331.AD7BEEC4@sybase.com>
Date: Sun, 13 Jan 2002 10:26:09 -0700
From: Bret Halford <bret@sybase.com>
Organization: Sybase, Inc.
X-Mailer: Mozilla 4.76 [en] (Windows NT 5.0; U)
X-Accept-Language: en
MIME-Version: 1.0
Subject: Re: high-performance wish list: no logging option for a device
References: <MvxQ#n7mBHA.297@forums.sybase.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Newsgroups: sybase.public.ase.product_futures_discussion
Lines: 18
NNTP-Posting-Host: 130.214.8.36
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:938
Article PK: 95176


Pablo Sanchez wrote:

> Howdy!
>
> I'd like to suggest that Sybase Engineering offer the ability to
> create and alter devices so logging is off or on per device. This
> would allow me set aside certain devices with no logging set, and
> assign them to say, tempdb.
>
> It might be a misnomer to say 'no logging' in that we might wish to
> perform some minimal logging.
>
>

Hi Pablo,

Minimal logging of allocations would still be necessary, I'm sure. What
do you
propose happens to a transaction if it hits a fatal error or any other
condition
that normally requires a rollback? Should ASE simply commit the changes
it made so far?

-bret


Pablo Sanchez Posted on 2002-01-13 22:53:30.0Z
From: "Pablo Sanchez" <pablo@dev.null>
References: <MvxQ#n7mBHA.297@forums.sybase.com> <3C41C331.AD7BEEC4@sybase.com>
Subject: Re: high-performance wish list: no logging option for a device
Date: Sun, 13 Jan 2002 15:53:30 -0700
Lines: 56
Organization: High-Performance Database Engineering
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <ChPZSXInBHA.190@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 207.225.105.222
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:936
Article PK: 95177


"Bret Halford" <bret@sybase.com> wrote in message
news:3C41C331.AD7BEEC4@sybase.com...
>
>
> Pablo Sanchez wrote:
>
> > Howdy!
> >
> > I'd like to suggest that Sybase Engineering offer the ability to
> > create and alter devices so logging is off or on per device. This
> > would allow me set aside certain devices with no logging set, and
> > assign them to say, tempdb.
> >
> > It might be a misnomer to say 'no logging' in that we might wish
to
> > perform some minimal logging.
> >
> >
>
> Hi Pablo,
>
> Minimal logging of allocations would still be necessary, I'm sure.
What
> do you
> propose happens to a transaction if it hits a fatal error or any
other
> condition
> that normally requires a rollback? Should ASE simply commit the
changes
> it made so far?

Hey Bret,

A little bit more refinement on the idea ... we might either handle
this at the device level or at the DB-level (sp_dboption). I kinda
like the DB level but understand that switching the option from on to
off would more at least require a database dump to ensure
recoverability. Thinking aloud, it might provide easier user
flexibility to have the option at the DB-level.

To answer you question (I'm a windbag I guess!), I propose that the
definition of no logging means that all transactions are treated as
single transactions. In other words, if I have a transaction that is
comprised of multiple DML statements, these degenerate to single
transactions when the DB is set to 'no logging.'

If any one DML fail, processing stops. There will be no rollback.
The data may be inconsistent but that's the given with 'no logging'
--
Pablo Sanchez, High-Performance Database Engineering
www.hpdbe.com
Independent Contractor, available for short-term and long-term
contracts


Sethu Posted on 2002-03-05 02:46:58.0Z
Message-ID: <3C8431A2.A4D34B96@sybase.com>
Date: Mon, 04 Mar 2002 21:46:58 -0500
From: Sethu <sethu@sybase.com>
Organization: Sybase, Inc.
X-Mailer: Mozilla 4.79 [en] (Windows NT 5.0; U)
X-Accept-Language: en
MIME-Version: 1.0
Subject: Re: high-performance wish list: no logging option for a device
References: <MvxQ#n7mBHA.297@forums.sybase.com> <3C41C331.AD7BEEC4@sybase.com> <ChPZSXInBHA.190@forums.sybase.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Newsgroups: sybase.public.ase.product_futures_discussion
Lines: 29
NNTP-Posting-Host: 10.22.91.118
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:750
Article PK: 94276

Hi,

How about some restrictions like existing data rows won't be expanded upon
update?

Sethu

> Hey Bret,
>
> A little bit more refinement on the idea ... we might either handle
> this at the device level or at the DB-level (sp_dboption). I kinda
> like the DB level but understand that switching the option from on to
> off would more at least require a database dump to ensure
> recoverability. Thinking aloud, it might provide easier user
> flexibility to have the option at the DB-level.
>
> To answer you question (I'm a windbag I guess!), I propose that the
> definition of no logging means that all transactions are treated as
> single transactions. In other words, if I have a transaction that is
> comprised of multiple DML statements, these degenerate to single
> transactions when the DB is set to 'no logging.'
>
> If any one DML fail, processing stops. There will be no rollback.
> The data may be inconsistent but that's the given with 'no logging'
> --
> Pablo Sanchez, High-Performance Database Engineering
> www.hpdbe.com
> Independent Contractor, available for short-term and long-term
> contracts


Pablo Sanchez Posted on 2002-03-08 00:02:33.0Z
From: "Pablo Sanchez" <pablo@dev.null>
References: <MvxQ#n7mBHA.297@forums.sybase.com> <3C41C331.AD7BEEC4@sybase.com> <ChPZSXInBHA.190@forums.sybase.com> <3C8431A2.A4D34B96@sybase.com>
Subject: Re: high-performance wish list: no logging option for a device
Date: Thu, 7 Mar 2002 17:02:33 -0700
Lines: 19
Organization: High-Performance Database Engineering
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <17QFFWjxBHA.332@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 207.225.105.222
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:735
Article PK: 94262


"Sethu" <sethu@sybase.com> wrote in message
news:3C8431A2.A4D34B96@sybase.com...
> Hi,
>
> How about some restrictions like existing data rows won't be
expanded upon
> update?

For my own edification, why would we place the above restriction? I
realize that we'd have to log some of this data. What I envision is
creating an equivalent /dev/null in the ASE kernel. The outer layer
codes would 'log' but it'd go to the old bit bucket.
--
Pablo Sanchez, High-Performance Database Engineering
www.hpdbe.com
Available for short-term and long-term contracts


Sethu Posted on 2002-03-09 03:09:51.0Z
Message-ID: <3C897CFF.BD089661@sybase.com>
Date: Fri, 08 Mar 2002 19:09:51 -0800
From: Sethu <sethu@sybase.com>
Organization: Sybase, Inc.
X-Mailer: Mozilla 4.79 [en] (Windows NT 5.0; U)
X-Accept-Language: en
MIME-Version: 1.0
Subject: Re: high-performance wish list: no logging option for a device
References: <MvxQ#n7mBHA.297@forums.sybase.com> <3C41C331.AD7BEEC4@sybase.com> <ChPZSXInBHA.190@forums.sybase.com> <3C8431A2.A4D34B96@sybase.com> <17QFFWjxBHA.332@forums.sybase.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Newsgroups: sybase.public.ase.product_futures_discussion
Lines: 19
NNTP-Posting-Host: 10.22.91.118
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:715
Article PK: 94242

The reason is that in cases where rows have to expand,
ASE does deferred updates. That will incur logging.

Sethu

Pablo Sanchez wrote:
>
> "Sethu" <sethu@sybase.com> wrote in message
> news:3C8431A2.A4D34B96@sybase.com...
> > Hi,
> >
> > How about some restrictions like existing data rows won't be
> expanded upon
> > update?
>
> For my own edification, why would we place the above restriction? I
> realize that we'd have to log some of this data. What I envision is
> creating an equivalent /dev/null in the ASE kernel. The outer layer
> codes would 'log' but it'd go to the old bit bucket.
> --
> Pablo Sanchez, High-Performance Database Engineering
> www.hpdbe.com
> Available for short-term and long-term contracts


Pablo Sanchez Posted on 2002-03-09 03:53:50.0Z
From: "Pablo Sanchez" <pablo@dev.null>
References: <MvxQ#n7mBHA.297@forums.sybase.com> <3C41C331.AD7BEEC4@sybase.com> <ChPZSXInBHA.190@forums.sybase.com> <3C8431A2.A4D34B96@sybase.com> <17QFFWjxBHA.332@forums.sybase.com> <3C897CFF.BD089661@sybase.com>
Subject: Re: high-performance wish list: no logging option for a device
Date: Fri, 8 Mar 2002 20:53:50 -0700
Lines: 22
Organization: High-Performance Database Engineering
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <8z1PA8xxBHA.311@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 207.225.105.222
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:713
Article PK: 94244


"Sethu" <sethu@sybase.com> wrote in message
news:3C897CFF.BD089661@sybase.com...
> The reason is that in cases where rows have to expand,
> ASE does deferred updates. That will incur logging.

Cool. The way I envisioned ASE handling this option was as I
previously noted: log = /dev/null. The deferred update would call
the underlying routines to log, however, those underlying routines
would recognize that no logging is to be done, therefore would return
without actually writing log info.

I know that I'm being simplistic but I believe ASE is properly
abstracted.

Thx!
--
Pablo Sanchez, High-Performance Database Engineering
www.hpdbe.com
Available for short-term and long-term contracts


Rob Verschoor Posted on 2002-03-09 15:44:04.0Z
Reply-To: "Rob Verschoor" <rob@DO.NOT.SPAM.sypron.nl>
From: "Rob Verschoor" <rob@DO.NOT.SPAM.sypron.nl>
References: <MvxQ#n7mBHA.297@forums.sybase.com> <3C41C331.AD7BEEC4@sybase.com> <ChPZSXInBHA.190@forums.sybase.com> <3C8431A2.A4D34B96@sybase.com> <17QFFWjxBHA.332@forums.sybase.com> <3C897CFF.BD089661@sybase.com> <8z1PA8xxBHA.311@forums.sybase.com>
Subject: Re: high-performance wish list: no logging option for a device
Date: Sat, 9 Mar 2002 16:44:04 +0100
Lines: 37
Organization: Sypron B.V.
MIME-Version: 1.0
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2919.6600
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
Message-ID: <$Wkz9J4xBHA.214@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: asd3-p2154.dial.euronet.nl 81.68.8.106
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:710
Article PK: 94237

The basic issue with deferred updates is that the stuff written to the
log must be read back to complete the update (the log is used as a
kind of "scratch pad" if you like). Therefore, simply not doing any
logging is too simple, as this will impact certain functionality.
I suppose it's impossible to do any deferrred updates/deletes without
logging, so I would expect such updates/deletes to result in run-time
errors -- and I'd be perfectly happy with that if it allowed me to do
unlogged DML.

Rob V.

"Pablo Sanchez" <pablo@dev.null> wrote in message
news:8z1PA8xxBHA.311@forums.sybase.com...
>
> "Sethu" <sethu@sybase.com> wrote in message
> news:3C897CFF.BD089661@sybase.com...
> > The reason is that in cases where rows have to expand,
> > ASE does deferred updates. That will incur logging.
>
> Cool. The way I envisioned ASE handling this option was as I
> previously noted: log = /dev/null. The deferred update would call
> the underlying routines to log, however, those underlying routines
> would recognize that no logging is to be done, therefore would
return
> without actually writing log info.
>
> I know that I'm being simplistic but I believe ASE is properly
> abstracted.
>
> Thx!
> --
> Pablo Sanchez, High-Performance Database Engineering
> www.hpdbe.com
> Available for short-term and long-term contracts
>
>


Marc Zampetti Posted on 2002-03-28 15:39:48.0Z
Message-ID: <3CA33944.5000605@aol.com>
Date: Thu, 28 Mar 2002 10:39:48 -0500
From: Marc Zampetti <zampmarc@aol.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.4) Gecko/20011128 Netscape6/6.2.1
X-Accept-Language: en-us
MIME-Version: 1.0
Subject: Re: high-performance wish list: no logging option for a device
References: <MvxQ#n7mBHA.297@forums.sybase.com> <3C41C331.AD7BEEC4@sybase.com> <ChPZSXInBHA.190@forums.sybase.com> <3C8431A2.A4D34B96@sybase.com> <17QFFWjxBHA.332@forums.sybase.com> <3C897CFF.BD089661@sybase.com> <8z1PA8xxBHA.311@forums.sybase.com> <$Wkz9J4xBHA.214@forums.sybase.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Newsgroups: sybase.public.ase.product_futures_discussion
Lines: 52
NNTP-Posting-Host: pix-fw.wan.aol.com 152.163.190.1
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:593
Article PK: 94127


Rob Verschoor wrote:

> The basic issue with deferred updates is that the stuff written to the
> log must be read back to complete the update (the log is used as a
> kind of "scratch pad" if you like). Therefore, simply not doing any
> logging is too simple, as this will impact certain functionality.
> I suppose it's impossible to do any deferrred updates/deletes without
> logging, so I would expect such updates/deletes to result in run-time
> errors -- and I'd be perfectly happy with that if it allowed me to do
> unlogged DML.
>
> Rob V.
>
> "Pablo Sanchez" <pablo@dev.null> wrote in message
> news:8z1PA8xxBHA.311@forums.sybase.com...
>
>>"Sethu" <sethu@sybase.com> wrote in message
>>news:3C897CFF.BD089661@sybase.com...
>>
>>>The reason is that in cases where rows have to expand,
>>>ASE does deferred updates. That will incur logging.
>>>
>>Cool. The way I envisioned ASE handling this option was as I
>>previously noted: log = /dev/null. The deferred update would call
>>the underlying routines to log, however, those underlying routines
>>would recognize that no logging is to be done, therefore would
>>
> return
>
>>without actually writing log info.
>>
>>I know that I'm being simplistic but I believe ASE is properly
>>abstracted.
>>
>>Thx!
>>--
>>Pablo Sanchez, High-Performance Database Engineering
>>www.hpdbe.com
>>Available for short-term and long-term contracts
>>
>>
>>
>

I think we are getting too complicated here. The real issue is turn off
the persistent aspect of the transaction log. I have a high-volumne,
volatile data application, where if I loose some data, then so be it. I
simply want to store the data in the database because RDBMS are very
good at handling millions of discreet objects. For me, make the tran log
in memory, and never write to disk is what I want. In the case of
orderly shutdowns, flushing the buffer cache to disk ensures a consitent
database. In other cases, simply document the risks (like Sybase does
for uses of buffered files).

Marc