Re: IFX_FOLDVIEW , vantages? disvantages?

Posted in 2010

Topics: Performance & Tuning, Triggers, Constraints & Referential Integrity

Fernando Nunes — Fri, 08 Oct 2010 14:17:27 — source: IIUG Forums & Mailing Lists

On Fri, Oct 8, 2010 at 12:07 PM, Cesar Inacio Martins < cesar_inacio_martins@yahoo.com.br> wrote: > Hi, > > By default the IFX_FOLDVIEW is disabled . > > I read in some places of the documentation about the vantages to active > this option. But no disvantages are mentioned... or risks of degrade the > performance when enable it, or others consequences (problem with triggers? > permission?) > > So, anyone can explain to me , why not keep this option enabled by default > ? > > I don't see any disadvantage. But this raises the question: Why the hell does the parameter exist? My *personal* opinion on this is that this and other parameters exist because R&D cannot be 110% sure that the new behavior will not be bad at times. So the parameter allows the engine to fallback to the "old" behavior. I recall having hit a bug with this activated but it was on the first fixpack where it was introduced. There are other parameters that follow the same idea: BATCHEDREAD_TABLE, INDEX_SJ and we may see others in the future. Although I could easily disagree with their existence, I understand they are important. They allow us to introduce new features while give control to the users to not use them in case they found problems. Ideally, new features would not have bugs and we would never need these kind of parameters, but we live in real world and not on a perfect one :) This parameter specifically change the behavior of the optimizer. And this can be considered a high risk activity... So, my advice: Turn it on on less critical systems and wait... If you don't have problems, turn it on production. I don't recall exactly if it was introduced in 11.10 or 11.50, but in any case the bugs should be gone in latests fixpacks (don't recall seeing any fix regarding this recently). Regards. > _______________________________________________ > Informix-list mailing list > Informix-list@iiug.org > http://www.iiug.org/mailman/listinfo/informix-list > > -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently... --00163645906a562a1204921bab77

↪ replying to Fernando Nunes

Peter_Logan@spartanstores.com — Fri, 08 Oct 2010 14:23:24 — source: IIUG Forums & Mailing Lists

Is there an environment variable that goes with this ... so that you can test it on a session ... before turning it on the entire instance ??? Peter Logan Senior Database Administrator Phone: 616/878-8309 From: "Fernando Nunes" <domusonline@gmail.com> To: ids@iiug.org Date: 10/08/2010 10:18 AM Subject: Re: IFX_FOLDVIEW , vantages? disvantages? [21602] Sent by: ids-bounces@iiug.org On Fri, Oct 8, 2010 at 12:07 PM, Cesar Inacio Martins < cesar_inacio_martins@yahoo.com.br> wrote: > Hi, > > By default the IFX_FOLDVIEW is disabled . > > I read in some places of the documentation about the vantages to active > this option. But no disvantages are mentioned... or risks of degrade the > performance when enable it, or others consequences (problem with triggers? > permission?) > > So, anyone can explain to me , why not keep this option enabled by default > ? > > I don't see any disadvantage. But this raises the question: Why the hell does the parameter exist? My *personal* opinion on this is that this and other parameters exist because R&D cannot be 110% sure that the new behavior will not be bad at times. So the parameter allows the engine to fallback to the "old" behavior. I recall having hit a bug with this activated but it was on the first fixpack where it was introduced. There are other parameters that follow the same idea: BATCHEDREAD_TABLE, INDEX_SJ and we may see others in the future. Although I could easily disagree with their existence, I understand they are important. They allow us to introduce new features while give control to the users to not use them in case they found problems. Ideally, new features would not have bugs and we would never need these kind of parameters, but we live in real world and not on a perfect one :) This parameter specifically change the behavior of the optimizer. And this can be considered a high risk activity... So, my advice: Turn it on on less critical systems and wait... If you don't have problems, turn it on production. I don't recall exactly if it was introduced in 11.10 or 11.50, but in any case the bugs should be gone in latests fixpacks (don't recall seeing any fix regarding this recently). Regards. > _______________________________________________ > Informix-list mailing list > Informix-list@iiug.org > http://www.iiug.org/mailman/listinfo/informix-list > > -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently... --00163645906a562a1204921bab77 ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.

↪ replying to Peter_Logan@spartanstores.com

Art Kagel — Fri, 08 Oct 2010 14:43:10 — source: IIUG Forums & Mailing Lists

I agree with Fernando that parameters like these that allow us to disable new features are a very good thing. When they decide to make the default enabled or disabled often has to do with what happened during alpha level testing with 3rd party packages. Some software that made extensive use of views probably had some performance problems due to inappropriate view use and they were working to fix the problem in their next release. To avoid customers with that/those packages having instant performance problems the default was made to be disabled in this case. My own testing and usage at customer sites mainly supports turning this one one. However, keeping with my theme for 2010 that Good DBAs are Paranoid DBAs, I would definitely test this on a development machine and a Q&A machine. Better yet, if you can do some definable and repeatable performance tests using a home grown tool or a commercial one like iWatch to prove the affect of turning this on, do it. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) IIUG Board of Directors (art@iiug.org) Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Fri, Oct 8, 2010 at 10:23 AM, Peter_Logan@spartanstores.com < Peter_Logan@spartanstores.com> wrote: > Is there an environment variable that goes with this ... so that you can > test it on a session ... before turning it on the entire instance ??? > > Peter Logan > Senior Database Administrator > Phone: 616/878-8309 > > From: > "Fernando Nunes" <domusonline@gmail.com> > To: > ids@iiug.org > Date: > 10/08/2010 10:18 AM > Subject: > Re: IFX_FOLDVIEW , vantages? disvantages? [21602] > Sent by: > ids-bounces@iiug.org > > On Fri, Oct 8, 2010 at 12:07 PM, Cesar Inacio Martins < > cesar_inacio_martins@yahoo.com.br> wrote: > > > Hi, > > > > By default the IFX_FOLDVIEW is disabled . > > > > I read in some places of the documentation about the vantages to active > > this option. But no disvantages are mentioned... or risks of degrade the > > > performance when enable it, or others consequences (problem with > triggers? > > permission?) > > > > So, anyone can explain to me , why not keep this option enabled by > default > > ? > > > > > I don't see any disadvantage. But this raises the question: Why the hell > does the parameter exist? > My *personal* opinion on this is that this and other parameters exist > because R&D cannot be 110% sure that the new behavior will not be bad at > times. So the parameter allows the engine to fallback to the "old" > behavior. > > I recall having hit a bug with this activated but it was on the first > fixpack where it was introduced. > > There are other parameters that follow the same idea: BATCHEDREAD_TABLE, > INDEX_SJ and we may see others in the future. Although I could easily > disagree with their existence, I understand they are important. They allow > > us to introduce new features while give control to the users to not use > them > in case they found problems. Ideally, new features would not have bugs and > > we would never need these kind of parameters, but we live in real world > and > not on a perfect one :) > > This parameter specifically change the behavior of the optimizer. And this > > can be considered a high risk activity... > > So, my advice: Turn it on on less critical systems and wait... If you > don't > have problems, turn it on production. > I don't recall exactly if it was introduced in 11.10 or 11.50, but in any > case the bugs should be gone in latests fixpacks (don't recall seeing any > fix regarding this recently). > > Regards. > > > _______________________________________________ > > Informix-list mailing list > > Informix-list@iiug.org > > http://www.iiug.org/mailman/listinfo/informix-list > > > > > > -- > Fernando Nunes > Portugal > > http://informix-technology.blogspot.com > My email works... but I don't check it frequently... > > --00163645906a562a1204921bab77 > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --00163628370841518e04921c07bf

↪ replying to Fernando Nunes

Cesar Inacio Martins — Fri, 08 Oct 2010 15:05:48 — source: IIUG Forums & Mailing Lists

Thank's Fernando! --- Em sex, 8/10/10, Fernando Nunes <domusonline@gmail.com> escreveu: De: Fernando Nunes <domusonline@gmail.com> Assunto: Re: IFX_FOLDVIEW , vantages? disvantages? [21602] Para: ids@iiug.org Data: Sexta-feira, 8 de Outubro de 2010, 11:17 On Fri, Oct 8, 2010 at 12:07 PM, Cesar Inacio Martins < cesar_inacio_martins@yahoo.com.br> wrote: > Hi, > > By default the IFX_FOLDVIEW is disabled . > > I read in some places of the documentation about the vantages to active > this option. But no disvantages are mentioned... or risks of degrade the > performance when enable it, or others consequences (problem with triggers? > permission?) > > So, anyone can explain to me , why not keep this option enabled by default > ? > > I don't see any disadvantage. But this raises the question: Why the hell does the parameter exist? My *personal* opinion on this is that this and other parameters exist because R&D cannot be 110% sure that the new behavior will not be bad at times. So the parameter allows the engine to fallback to the "old" behavior. I recall having hit a bug with this activated but it was on the first fixpack where it was introduced. There are other parameters that follow the same idea: BATCHEDREAD_TABLE, INDEX_SJ and we may see others in the future. Although I could easily disagree with their existence, I understand they are important. They allow us to introduce new features while give control to the users to not use them in case they found problems. Ideally, new features would not have bugs and we would never need these kind of parameters, but we live in real world and not on a perfect one :) This parameter specifically change the behavior of the optimizer. And this can be considered a high risk activity... So, my advice: Turn it on on less critical systems and wait... If you don't have problems, turn it on production. I don't recall exactly if it was introduced in 11.10 or 11.50, but in any case the bugs should be gone in latests fixpacks (don't recall seeing any fix regarding this recently). Regards. > _______________________________________________ > Informix-list mailing list > Informix-list@iiug.org > http://www.iiug.org/mailman/listinfo/informix-list > > -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently... --00163645906a562a1204921bab77 ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.

↪ replying to Art Kagel

Cesar Inacio Martins — Fri, 08 Oct 2010 22:59:19 — source: IIUG Forums & Mailing Lists

Thanks for your answer Art, I will start to test it . Regards Cesar --- Em sex, 8/10/10, Art Kagel <art.kagel@gmail.com> escreveu: De: Art Kagel <art.kagel@gmail.com> Assunto: Re: IFX_FOLDVIEW , vantages? disvantages? [21604] Para: ids@iiug.org Data: Sexta-feira, 8 de Outubro de 2010, 11:43 I agree with Fernando that parameters like these that allow us to disable new features are a very good thing. When they decide to make the default enabled or disabled often has to do with what happened during alpha level testing with 3rd party packages. Some software that made extensive use of views probably had some performance problems due to inappropriate view use and they were working to fix the problem in their next release. To avoid customers with that/those packages having instant performance problems the default was made to be disabled in this case. My own testing and usage at customer sites mainly supports turning this one one. However, keeping with my theme for 2010 that Good DBAs are Paranoid DBAs, I would definitely test this on a development machine and a Q&A machine. Better yet, if you can do some definable and repeatable performance tests using a home grown tool or a commercial one like iWatch to prove the affect of turning this on, do it. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) IIUG Board of Directors (art@iiug.org) Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Fri, Oct 8, 2010 at 10:23 AM, Peter_Logan@spartanstores.com < Peter_Logan@spartanstores.com> wrote: > Is there an environment variable that goes with this ... so that you can > test it on a session ... before turning it on the entire instance ??? > > Peter Logan > Senior Database Administrator > Phone: 616/878-8309 > > From: > "Fernando Nunes" <domusonline@gmail.com> > To: > ids@iiug.org > Date: > 10/08/2010 10:18 AM > Subject: > Re: IFX_FOLDVIEW , vantages? disvantages? [21602] > Sent by: > ids-bounces@iiug.org > > On Fri, Oct 8, 2010 at 12:07 PM, Cesar Inacio Martins < > cesar_inacio_martins@yahoo.com.br> wrote: > > > Hi, > > > > By default the IFX_FOLDVIEW is disabled . > > > > I read in some places of the documentation about the vantages to active > > this option. But no disvantages are mentioned... or risks of degrade the > > > performance when enable it, or others consequences (problem with > triggers? > > permission?) > > > > So, anyone can explain to me , why not keep this option enabled by > default > > ? > > > > > I don't see any disadvantage. But this raises the question: Why the hell > does the parameter exist? > My *personal* opinion on this is that this and other parameters exist > because R&D cannot be 110% sure that the new behavior will not be bad at > times. So the parameter allows the engine to fallback to the "old" > behavior. > > I recall having hit a bug with this activated but it was on the first > fixpack where it was introduced. > > There are other parameters that follow the same idea: BATCHEDREAD_TABLE, > INDEX_SJ and we may see others in the future. Although I could easily > disagree with their existence, I understand they are important. They allow > > us to introduce new features while give control to the users to not use > them > in case they found problems. Ideally, new features would not have bugs and > > we would never need these kind of parameters, but we live in real world > and > not on a perfect one :) > > This parameter specifically change the behavior of the optimizer. And this > > can be considered a high risk activity... > > So, my advice: Turn it on on less critical systems and wait... If you > don't > have problems, turn it on production. > I don't recall exactly if it was introduced in 11.10 or 11.50, but in any > case the bugs should be gone in latests fixpacks (don't recall seeing any > fix regarding this recently). > > Regards. > > > _______________________________________________ > > Informix-list mailing list > > Informix-list@iiug.org > > http://www.iiug.org/mailman/listinfo/informix-list > > > > > > -- > Fernando Nunes > Portugal > > http://informix-technology.blogspot.com > My email works... but I don't check it frequently... > > --00163645906a562a1204921bab77 > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --00163628370841518e04921c07bf ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.