set up a secondary server tutorial info?
A user on IDS 12.10.FC9 under Windows Server 2012 R2 asked for a tutorial on building a failover secondary server, plus questions about matching OS/IDS versions and upgrades. Replies suggested HDR or RSS (possibly with DELAY_APPLY), or continuous log recovery if the standby must not serve queries; same major version and fixpack level is recommended, platforms should match (Server 2016 wasn't certified), and rolling upgrades can avoid re-instantiating the standby. Links to IBM developerWorks HA/cluster and rolling-upgrade articles and a personal HDR tutorial were given, plus an IIUG conference-papers pointer. The thread then drifted into a Windows-vs-Linux performance debate; no confirmed outcome is recorded, as the poster planned to work through the links later.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
IDS 12.10.FC9 (workgroup edition)
Windows Server 2012 r2
Hi,
We have one informix instance for our production database system.
I'm about to set up a secondary server for the first time. This will be for
failover should the primary server fail, but will be unavailable for sql
read/write queries to get the benefit of not needing more informix licenses.
It will be on the same network segment and kept in sync as near as practicable.
Does anyone know of a good tutorial document that covers everything I need to
know to accomplish this, and the follow on tasks of actually performing
failovers?
The administrators guide was confounding and left me unsure of how it all hung
together.
And questions related to this:
- not available for queries but able to be kept in sync - is that just via an
onmode option?
- Do the operating systems have to be the same, or can one be running Server
2012r2 while the other is on Server 2016?
- Do the IDS versions have to be the same, or just close enough?
- What happens when IDS is updated, do both primary and secondary need
updating together to keep synchronised? (this is especially relevant as it was
an update last year that borked our server for about 6 hours. This lead us to
think that having a secondary we could have fired up instead would have been a
good thing, but it would need to still have been on the previous version to
not suffer the same fate as the first server).
Regards,
Bryce Stenberg
Harness Racing New Zealand Inc.
Suggestion 1 - upgrade to linux. Performance is 10x on the same hardware.
Failing that, find Guy Bowerman and check with him. He is the Informix/MSSoft
master.
j.
> On Jan 25, 2018, at 10:29 PM, BRYCE STENBERG <bryce@hrnz.co.nz> wrote:
>
> IDS 12.10.FC9 (workgroup edition)
> Windows Server 2012 r2
>
> Hi,
>
> We have one informix instance for our production database system.
> I'm about to set up a secondary server for the first time. This will be for
> failover should the primary server fail, but will be unavailable for sql
> read/write queries to get the benefit of not needing more informix licenses.
> It will be on the same network segment and kept in sync as near as
> practicable.
>
> Does anyone know of a good tutorial document that covers everything I need to
> know to accomplish this, and the follow on tasks of actually performing
> failovers?
> The administrators guide was confounding and left me unsure of how it all
hung
> together.
>
> And questions related to this:
> - not available for queries but able to be kept in sync - is that just via an
> onmode option?
> - Do the operating systems have to be the same, or can one be running Server
> 2012r2 while the other is on Server 2016?
> - Do the IDS versions have to be the same, or just close enough?
> - What happens when IDS is updated, do both primary and secondary need
> updating together to keep synchronised? (this is especially relevant as it
was
> an update last year that borked our server for about 6 hours. This lead us to
> think that having a secondary we could have fired up instead would have been
a
> good thing, but it would need to still have been on the previous version to
> not suffer the same fate as the first server).
>
> Regards,
> Bryce Stenberg
> Harness Racing New Zealand Inc.
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
Cheers Jack, I appreciate linux is faster but even with linux I would still be looking for this tutorial info... Regards, Bryce.
First of all: ignore the usual Windows bashing - I'd see no reason to
believe Informix would be slower or faster than with a different OS on same
HW, provided proper handling. How nicely Informix integrates into Windows
is a separate question, but shouldn't affect performance.
HDR probably is what you're aiming at, maybe RSS (more asynchronous, less
impacting on production). It's a bit more complicated to set up and get
going on Windows, but it is being used.
'Same platform' is advised, esp. if same load should be able to run on
alternate system - in case of fail-over. Exact same OS isn't a must, yet
Server 2016, afaik, isn't certified yet for Informix (should you find it to
be doing well with Informix, I'd not object to using it).
There's not really a 'not available for queries but able to be kept in sync
' mode with any of Informix secondary types. You might want to look into
'continuous log recovery' instead: a restored clone of your production,
kept in permanent recovery mode applying production logical logs as they
are being shipped from production. This of course wouldn't be as close in
sync as any replication type.
Re. IDS versions: recommendation is 'same' - what reasons would you have
for 'differing'? Certainly major version (11.50, 11.70, 12.10) needs to be
same, then there are fixpack levels that would play well with each other,
others less well.
Then the upgrade question: you'd definitely be better off with two
systems, even if the standby had to be re-instantiated after the upgrade,
and with recent versions even that requirement could be avoided.
HTH,
Andreas
From: "BRYCE STENBERG" <bryce@hrnz.co.nz>
To: ids@iiug.org
Date: 01/26/2018 04:30 AM
Subject: set up a secondary server tutorial info? [40576]
Sent by: ids-bounces@iiug.org
IDS 12.10.FC9 (workgroup edition)
Windows Server 2012 r2
Hi,
We have one informix instance for our production database system.
I'm about to set up a secondary server for the first time. This will be for
failover should the primary server fail, but will be unavailable for sql
read/write queries to get the benefit of not needing more informix
licenses.
It will be on the same network segment and kept in sync as near as
practicable.
Does anyone know of a good tutorial document that covers everything I need
to
know to accomplish this, and the follow on tasks of actually performing
failovers?
The administrators guide was confounding and left me unsure of how it all
hung
together.
And questions related to this:
- not available for queries but able to be kept in sync - is that just via
an
onmode option?
- Do the operating systems have to be the same, or can one be running
Server
2012r2 while the other is on Server 2016?
- Do the IDS versions have to be the same, or just close enough?
- What happens when IDS is updated, do both primary and secondary need
updating together to keep synchronised? (this is especially relevant as it
was
an update last year that borked our server for about 6 hours. This lead us
to
think that having a secondary we could have fired up instead would have
been a
good thing, but it would need to still have been on the previous version to
not suffer the same fate as the first server).
Regards,
Bryce Stenberg
Harness Racing New Zealand Inc.
*******************************************************************************
Forum Note: Use "Reply" to post a response in the discussion forum.
Hi, Bryce.
Just found this DeveloperWorks article.
It is outdated, but the basics are the same, maybe it can give you a clear
idea of how Informix deals with HA/failover on a cluster.
https://www.ibm.com/developerworks/data/roadmaps/roadmap_ifx_hacluster.html
This is a newer one, about rolling window upgrades (cluster upgrades with no
downtime, a great feature).
https://www.ibm.com/developerworks/data/library/techarticle/dm-1012rollingupgrad
e/index.html
Hope it helps.
Best regards.
Alexandre Marini
IBM Informix Certified Professional v10 / v11.50 / v11.70 / v12.10
IBM Informix on Cloud - Database Administrator - 2017
IBM dashDB Managed Service for Analytics and Transactions - 2017
DB2 Advanced DBA - v10.5 for LUW
IBM Information Management Informix Technical Professional
IBM Certified Developer - Informix Genero
Informix independent consultant
________________________________
De: ids-bounces@iiug.org <ids-bounces@iiug.org> em nome de Andreas Legner1
<Andreas.Legner1@de.ibm.com>
Enviado: sexta-feira, 26 de janeiro de 2018 07:59
Para: ids@iiug.org
Assunto: Re: set up a secondary server tutorial info? [40580]
First of all: ignore the usual Windows bashing - I'd see no reason to
believe Informix would be slower or faster than with a different OS on same
HW, provided proper handling. How nicely Informix integrates into Windows
is a separate question, but shouldn't affect performance.
HDR probably is what you're aiming at, maybe RSS (more asynchronous, less
impacting on production). It's a bit more complicated to set up and get
going on Windows, but it is being used.
'Same platform' is advised, esp. if same load should be able to run on
alternate system - in case of fail-over. Exact same OS isn't a must, yet
Server 2016, afaik, isn't certified yet for Informix (should you find it to
be doing well with Informix, I'd not object to using it).
There's not really a 'not available for queries but able to be kept in sync
' mode with any of Informix secondary types. You might want to look into
'continuous log recovery' instead: a restored clone of your production,
kept in permanent recovery mode applying production logical logs as they
are being shipped from production. This of course wouldn't be as close in
sync as any replication type.
Re. IDS versions: recommendation is 'same' - what reasons would you have
for 'differing'? Certainly major version (11.50, 11.70, 12.10) needs to be
same, then there are fixpack levels that would play well with each other,
others less well.
Then the upgrade question: you'd definitely be better off with two
systems, even if the standby had to be re-instantiated after the upgrade,
and with recent versions even that requirement could be avoided.
HTH,
Andreas
From: "BRYCE STENBERG" <bryce@hrnz.co.nz>
To: ids@iiug.org
Date: 01/26/2018 04:30 AM
Subject: set up a secondary server tutorial info? [40576]
Sent by: ids-bounces@iiug.org
IDS 12.10.FC9 (workgroup edition)
Windows Server 2012 r2
Hi,
We have one informix instance for our production database system.
I'm about to set up a secondary server for the first time. This will be for
failover should the primary server fail, but will be unavailable for sql
read/write queries to get the benefit of not needing more informix
licenses.
It will be on the same network segment and kept in sync as near as
practicable.
Does anyone know of a good tutorial document that covers everything I need
to
know to accomplish this, and the follow on tasks of actually performing
failovers?
The administrators guide was confounding and left me unsure of how it all
hung
together.
And questions related to this:
- not available for queries but able to be kept in sync - is that just via
an
onmode option?
- Do the operating systems have to be the same, or can one be running
Server
2012r2 while the other is on Server 2016?
- Do the IDS versions have to be the same, or just close enough?
- What happens when IDS is updated, do both primary and secondary need
updating together to keep synchronised? (this is especially relevant as it
was
an update last year that borked our server for about 6 hours. This lead us
to
think that having a secondary we could have fired up instead would have
been a
good thing, but it would need to still have been on the previous version to
not suffer the same fate as the first server).
Regards,
Bryce Stenberg
Harness Racing New Zealand Inc.
*******************************************************************************
Forum Note: Use "Reply" to post a response in the discussion forum.
*******************************************************************************
Forum Note: Use "Reply" to post a response in the discussion forum.
I have a simple tutorial for HDR at
http://smooth1.co.uk/installs/dbinstalls.html#3.2.2
You could also have an RSS with DELAY_APPLY set.
I will write a tutorial for that over the weekend and improve the one above!
Regards,
David.
> On 26 January 2018 at 11:11 Alexandre Marini <alexandre_marini@hotmail.com>
wrote:
>
>
> Hi, Bryce.
> Just found this DeveloperWorks article.
> It is outdated, but the basics are the same, maybe it can give you a clear
> idea of how Informix deals with HA/failover on a cluster.
>
> https://www.ibm.com/developerworks/data/roadmaps/roadmap_ifx_hacluster.html
>
> This is a newer one, about rolling window upgrades (cluster upgrades with no
> downtime, a great feature).
>
>
>
https://www.ibm.com/developerworks/data/library/techarticle/dm-1012rollingupgrad
e/index.html
>
> Hope it helps.
> Best regards.
>
> Alexandre Marini
> IBM Informix Certified Professional v10 / v11.50 / v11.70 / v12.10
> IBM Informix on Cloud - Database Administrator - 2017
> IBM dashDB Managed Service for Analytics and Transactions - 2017
> DB2 Advanced DBA - v10.5 for LUW
> IBM Information Management Informix Technical Professional
> IBM Certified Developer - Informix Genero
> Informix independent consultant
> ________________________________
> De: ids-bounces@iiug.org <ids-bounces@iiug.org> em nome de Andreas Legner1
> <Andreas.Legner1@de.ibm.com>
> Enviado: sexta-feira, 26 de janeiro de 2018 07:59
> Para: ids@iiug.org
> Assunto: Re: set up a secondary server tutorial info? [40580]
>
> First of all: ignore the usual Windows bashing - I'd see no reason to
> believe Informix would be slower or faster than with a different OS on same
> HW, provided proper handling. How nicely Informix integrates into Windows
> is a separate question, but shouldn't affect performance.
>
> HDR probably is what you're aiming at, maybe RSS (more asynchronous, less
> impacting on production). It's a bit more complicated to set up and get
> going on Windows, but it is being used.
>
> 'Same platform' is advised, esp. if same load should be able to run on
> alternate system - in case of fail-over. Exact same OS isn't a must, yet
> Server 2016, afaik, isn't certified yet for Informix (should you find it to
> be doing well with Informix, I'd not object to using it).
>
> There's not really a 'not available for queries but able to be kept in sync
> ' mode with any of Informix secondary types. You might want to look into
> 'continuous log recovery' instead: a restored clone of your production,
> kept in permanent recovery mode applying production logical logs as they
> are being shipped from production. This of course wouldn't be as close in
> sync as any replication type.
>
> Re. IDS versions: recommendation is 'same' - what reasons would you have
> for 'differing'? Certainly major version (11.50, 11.70, 12.10) needs to be
> same, then there are fixpack levels that would play well with each other,
> others less well.
> Then the upgrade question: you'd definitely be better off with two
> systems, even if the standby had to be re-instantiated after the upgrade,
> and with recent versions even that requirement could be avoided.
>
> HTH,
> Andreas
>
> From: "BRYCE STENBERG" <bryce@hrnz.co.nz>
> To: ids@iiug.org
> Date: 01/26/2018 04:30 AM
> Subject: set up a secondary server tutorial info? [40576]
> Sent by: ids-bounces@iiug.org
>
> IDS 12.10.FC9 (workgroup edition)
> Windows Server 2012 r2
>
> Hi,
>
> We have one informix instance for our production database system.
> I'm about to set up a secondary server for the first time. This will be for
>
> failover should the primary server fail, but will be unavailable for sql
> read/write queries to get the benefit of not needing more informix
> licenses.
> It will be on the same network segment and kept in sync as near as
> practicable.
>
> Does anyone know of a good tutorial document that covers everything I need
> to
> know to accomplish this, and the follow on tasks of actually performing
> failovers?
> The administrators guide was confounding and left me unsure of how it all
> hung
> together.
>
> And questions related to this:
> - not available for queries but able to be kept in sync - is that just via
> an
> onmode option?
> - Do the operating systems have to be the same, or can one be running
> Server
> 2012r2 while the other is on Server 2016?
> - Do the IDS versions have to be the same, or just close enough?
> - What happens when IDS is updated, do both primary and secondary need
> updating together to keep synchronised? (this is especially relevant as it
> was
> an update last year that borked our server for about 6 hours. This lead us
> to
> think that having a secondary we could have fired up instead would have
> been a
> good thing, but it would need to still have been on the previous version to
>
> not suffer the same fate as the first server).
>
> Regards,
> Bryce Stenberg
> Harness Racing New Zealand Inc.
>
>
>
*******************************************************************************
>
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
Forgive my windows bashing. I just spent 11 years working with Informix on
Windows. As an idle experiment, I made a copy of the database onto a linux box
and fired the same application at both a Windows version and a Linux version
of the same database using the same data. The Linux version was 2x faster, the
Windows engine had been tuned, the Linux version - just thrown together.
Then there is the bit about Windows not fully understanding some Unix concepts
- like $TERM. Run a few queries through dbaccess - and oh yeah, sorry I dont
really know how to clear the screen, just read between the junk I left all
over it.
Load/unload - now that we have external tables its a lot better, but HPL and
Windows is not a thing, you were stuck with onpladm - which is fine, provided
you know everything that needs to be done. Oh, you want to unload to a
compression pipe? That was fun figuring out.
SMP - Windows does not play well here. Windows seems to use a clock, Oh, its
2:30 - Its time for processer #3 to handle the next thread - as opposed to a
good Linux SMP kernel which spreads the load evenly.
Memory management? That took a few weeks to work through.
The whole concept of the user Id Informix" - especially on a server in a
domain - this is painful. You can install without the Informix userid, oh -
you wanted to access the sysadmin database? ooops.
I could go on for a couple of hours here. If you are used to a Windows
environment and want to use Informix, that is one thing. If you have a Unix
background - dont torture yourself.
cheers
j.
> On Jan 26, 2018, at 4:59 AM, Andreas Legner1 <Andreas.Legner1@de.ibm.com>
wrote:
>
> First of all: ignore the usual Windows bashing - I'd see no reason to
> believe Informix would be slower or faster than with a different OS on same
> HW, provided proper handling. How nicely Informix integrates into Windows
> is a separate question, but shouldn't affect performance.
>
Did you try turning off the hyperthreading for Windows - I have seems that
make a difference
Paul Watson
Oninit www.oninit.com
Tel: +1 913 674 0360
Cell: +1 913 387 7529
Oninit® is a registered trademark of Oninit LLC
Failure is not as frightening as regret.
If you want to improve, be content to be thought foolish and stupid.
> On Jan 28, 2018, at 18:20, Jack Parker <jack.parker4@verizon.net> wrote:
>
> Forgive my windows bashing. I just spent 11 years working with Informix on
> Windows. As an idle experiment, I made a copy of the database onto a linux
box
> and fired the same application at both a Windows version and a Linux version
> of the same database using the same data. The Linux version was 2x faster,
the
> Windows engine had been tuned, the Linux version - just thrown together.
>
> Then there is the bit about Windows not fully understanding some Unix
concepts
> - like $TERM. Run a few queries through dbaccess - and oh yeah, sorry I dont
> really know how to clear the screen, just read between the junk I left all
> over it.
>
> Load/unload - now that we have external tables its a lot better, but HPL and
> Windows is not a thing, you were stuck with onpladm - which is fine, provided
> you know everything that needs to be done. Oh, you want to unload to a
> compression pipe? That was fun figuring out.
>
> SMP - Windows does not play well here. Windows seems to use a clock, Oh, its
> 2:30 - Its time for processer #3 to handle the next thread - as opposed to a
> good Linux SMP kernel which spreads the load evenly.
>
> Memory management? That took a few weeks to work through.
>
> The whole concept of the user Id Informix" - especially on a server in a
> domain - this is painful. You can install without the Informix userid, oh -
> you wanted to access the sysadmin database? ooops.
>
> I could go on for a couple of hours here. If you are used to a Windows
> environment and want to use Informix, that is one thing. If you have a Unix
> background - dont torture yourself.
>
> cheers
> j.
>
>> On Jan 26, 2018, at 4:59 AM, Andreas Legner1 <Andreas.Legner1@de.ibm.com>
> wrote:
>>
>> First of all: ignore the usual Windows bashing - I'd see no reason to
>> believe Informix would be slower or faster than with a different OS on same
>> HW, provided proper handling. How nicely Informix integrates into Windows
>> is a separate question, but shouldn't affect performance.
>>
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
--Apple-Mail-48352696-0954-4623-BDB7-8F1EA329D420
Yes, that was off.
I should have been clearer, the comparison between Windows and Linux was on
sister (identical) machines. Dont take my word for it, try it yourselves.
Im free of that employment now and can speak my mind to some degree - what,
you wonder where Ive been for the past decade?
j.
> On Jan 28, 2018, at 8:38 PM, Paul Watson <paul@oninit.com> wrote:
>
> Did you try turning off the hyperthreading for Windows - I have seems that
> make a difference
>
> Paul Watson
> Oninit www.oninit.com
> Tel: +1 913 674 0360
> Cell: +1 913 387 7529
>
> Oninit is a registered trademark of Oninit LLC
>
> Failure is not as frightening as regret.
> If you want to improve, be content to be thought foolish and stupid.
>
>> On Jan 28, 2018, at 18:20, Jack Parker <jack.parker4@verizon.net> wrote:
>>
>> Forgive my windows bashing. I just spent 11 years working with Informix on
>> Windows. As an idle experiment, I made a copy of the database onto a linux
> box
>> and fired the same application at both a Windows version and a Linux version
>> of the same database using the same data. The Linux version was 2x faster,
> the
>> Windows engine had been tuned, the Linux version - just thrown together.
>>
>> Then there is the bit about Windows not fully understanding some Unix
> concepts
>> - like $TERM. Run a few queries through dbaccess - and oh yeah, sorry I dont
>> really know how to clear the screen, just read between the junk I left all
>> over it.
>>
>> Load/unload - now that we have external tables its a lot better, but HPL and
>> Windows is not a thing, you were stuck with onpladm - which is fine,
> provided
>> you know everything that needs to be done. Oh, you want to unload to a
>> compression pipe? That was fun figuring out.
>>
>> SMP - Windows does not play well here. Windows seems to use a clock, Oh, its
>> 2:30 - Its time for processer #3 to handle the next thread - as opposed to a
>> good Linux SMP kernel which spreads the load evenly.
>>
>> Memory management? That took a few weeks to work through.
>>
>> The whole concept of the user Id Informix" - especially on a server in a
>> domain - this is painful. You can install without the Informix userid, oh -
>> you wanted to access the sysadmin database? ooops.
>>
>> I could go on for a couple of hours here. If you are used to a Windows
>> environment and want to use Informix, that is one thing. If you have a Unix
>> background - dont torture yourself.
>>
>> cheers
>> j.
>>
>>> On Jan 26, 2018, at 4:59 AM, Andreas Legner1 <Andreas.Legner1@de.ibm.com>
>> wrote:
>>>
>>> First of all: ignore the usual Windows bashing - I'd see no reason to
>>> believe Informix would be slower or faster than with a different OS on same
>>> HW, provided proper handling. How nicely Informix integrates into Windows
>>> is a separate question, but shouldn't affect performance.
>>>
>>
>>
>>
>
*******************************************************************************
>> Forum Note: Use "Reply" to post a response in the discussion forum.
>
> --Apple-Mail-48352696-0954-4623-BDB7-8F1EA329D420
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
I agree with you windows tends to be significantly slower for informix
Paul Watson
Oninit www.oninit.com
Tel: +1 913 674 0360
Cell: +1 913 387 7529
Oninit® is a registered trademark of Oninit LLC
Failure is not as frightening as regret.
If you want to improve, be content to be thought foolish and stupid.
> On Jan 28, 2018, at 20:21, Jack Parker <jack.parker4@verizon.net> wrote:
>
> Yes, that was off.
>
> I should have been clearer, the comparison between Windows and Linux was on
> sister (identical) machines. Dont take my word for it, try it yourselves.
>
> Im free of that employment now and can speak my mind to some degree - what,
> you wonder where Ive been for the past decade?
>
> j.
>
>> On Jan 28, 2018, at 8:38 PM, Paul Watson <paul@oninit.com> wrote:
>>
>> Did you try turning off the hyperthreading for Windows - I have seems that
>> make a difference
>>
>> Paul Watson
>> Oninit www.oninit.com
>> Tel: +1 913 674 0360
>> Cell: +1 913 387 7529
>>
>> Oninit is a registered trademark of Oninit LLC
>>
>> Failure is not as frightening as regret.
>> If you want to improve, be content to be thought foolish and stupid.
>>
>>> On Jan 28, 2018, at 18:20, Jack Parker <jack.parker4@verizon.net> wrote:
>>>
>>> Forgive my windows bashing. I just spent 11 years working with Informix on
>>> Windows. As an idle experiment, I made a copy of the database onto a linux
>> box
>>> and fired the same application at both a Windows version and a Linux
> version
>>> of the same database using the same data. The Linux version was 2x faster,
>> the
>>> Windows engine had been tuned, the Linux version - just thrown together.
>>>
>>> Then there is the bit about Windows not fully understanding some Unix
>> concepts
>>> - like $TERM. Run a few queries through dbaccess - and oh yeah, sorry I
> dont
>>> really know how to clear the screen, just read between the junk I left all
>>> over it.
>>>
>>> Load/unload - now that we have external tables its a lot better, but HPL
> and
>>> Windows is not a thing, you were stuck with onpladm - which is fine,
>> provided
>>> you know everything that needs to be done. Oh, you want to unload to a
>>> compression pipe? That was fun figuring out.
>>>
>>> SMP - Windows does not play well here. Windows seems to use a clock, Oh,
> its
>>> 2:30 - Its time for processer #3 to handle the next thread - as opposed to
> a
>>> good Linux SMP kernel which spreads the load evenly.
>>>
>>> Memory management? That took a few weeks to work through.
>>>
>>> The whole concept of the user Id Informix" - especially on a server in a
>>> domain - this is painful. You can install without the Informix userid, oh -
>>> you wanted to access the sysadmin database? ooops.
>>>
>>> I could go on for a couple of hours here. If you are used to a Windows
>>> environment and want to use Informix, that is one thing. If you have a Unix
>>> background - dont torture yourself.
>>>
>>> cheers
>>> j.
>>>
>>>> On Jan 26, 2018, at 4:59 AM, Andreas Legner1 <Andreas.Legner1@de.ibm.com>
>>> wrote:
>>>>
>>>> First of all: ignore the usual Windows bashing - I'd see no reason to
>>>> believe Informix would be slower or faster than with a different OS on
> same
>>>> HW, provided proper handling. How nicely Informix integrates into Windows
>>>> is a separate question, but shouldn't affect performance.
>>>>
>>>
>>>
>>>
>>
>
*******************************************************************************
>>> Forum Note: Use "Reply" to post a response in the discussion forum.
>>
>> --Apple-Mail-48352696-0954-4623-BDB7-8F1EA329D420
>>
>>
>>
>
*******************************************************************************
>> Forum Note: Use "Reply" to post a response in the discussion forum.
>>
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
--Apple-Mail-11B3E848-F0D5-4BBD-A34A-B99074E716A0
Hi,
Thanks everyone for the links regarding setting up a secondary server, going
to take me a week or two to work through them (due to annual leave now).
I was just curious about this 'hyperthreading off' making things faster for
Informix.
I have never seen this advice before, what is the reason it makes Informix run
faster? And is it only under windows, for Linux you leave hyperthreading
enabled in the bios because it uses it better?
What about everything else on the server that presumably benefits from
hyperthreading, like the OS and backup software for example?
(I should point out we only have 3 cores anyway due licence restrictions so
not much in the way of resources to begin with).
When I get back to work I'll try running some speed tests against our testing
server and see what results I get.
Regards,
Bryce Stenberg.
-----Original Message-----
From: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] On Behalf Of Jack
Parker
Sent: Monday, 29 January 2018 3:21 PM
To: ids@iiug.org
Subject: Re: set up a secondary server tutorial info? [40589]
Yes, that was off.
I should have been clearer, the comparison between Windows and Linux was on
sister (identical) machines. Dont take my word for it, try it yourselves.
Im free of that employment now and can speak my mind to some degree - what,
you wonder where Ive been for the past decade?
j.
> On Jan 28, 2018, at 8:38 PM, Paul Watson <paul@oninit.com> wrote:
>
> Did you try turning off the hyperthreading for Windows - I have seems
> that make a difference
>
> Paul Watson
> Oninit www.oninit.com
> Tel: +1 913 674 0360
> Cell: +1 913 387 7529
>
> Oninit is a registered trademark of Oninit LLC
>
> Failure is not as frightening as regret.
> If you want to improve, be content to be thought foolish and stupid.
>
>> On Jan 28, 2018, at 18:20, Jack Parker <jack.parker4@verizon.net> wrote:
>>
>> Forgive my windows bashing. I just spent 11 years working with
>> Informix on Windows. As an idle experiment, I made a copy of the
>> database onto a linux
> box
>> and fired the same application at both a Windows version and a Linux
version
>> of the same database using the same data. The Linux version was 2x
>> faster,
> the
>> Windows engine had been tuned, the Linux version - just thrown together.
>>
>> Then there is the bit about Windows not fully understanding some Unix
> concepts
>> - like $TERM. Run a few queries through dbaccess - and oh yeah, sorry
>> I
dont
>> really know how to clear the screen, just read between the junk I
>> left all over it.
>>
>> Load/unload - now that we have external tables its a lot better, but
>> HPL
and
>> Windows is not a thing, you were stuck with onpladm - which is fine,
> provided
>> you know everything that needs to be done. Oh, you want to unload to
>> a compression pipe? That was fun figuring out.
>>
>> SMP - Windows does not play well here. Windows seems to use a clock,
>> Oh,
its
>> 2:30 - Its time for processer #3 to handle the next thread - as
>> opposed to
a
>> good Linux SMP kernel which spreads the load evenly.
>>
>> Memory management? That took a few weeks to work through.
>>
>> The whole concept of the user Id Informix" - especially on a server
>> in a domain - this is painful. You can install without the Informix
>> userid, oh - you wanted to access the sysadmin database? ooops.
>>
>> I could go on for a couple of hours here. If you are used to a
>> Windows environment and want to use Informix, that is one thing. If
>> you have a Unix background - dont torture yourself.
>>
>> cheers
>> j.
>>
>>> On Jan 26, 2018, at 4:59 AM, Andreas Legner1
>>> <Andreas.Legner1@de.ibm.com>
>> wrote:
>>>
>>> First of all: ignore the usual Windows bashing - I'd see no reason
>>> to believe Informix would be slower or faster than with a different
>>> OS on
same
>>> HW, provided proper handling. How nicely Informix integrates into
>>> Windows is a separate question, but shouldn't affect performance.
>>>
>>
>>
>>
>
*******************************************************************************
>> Forum Note: Use "Reply" to post a response in the discussion forum.
>
> --Apple-Mail-48352696-0954-4623-BDB7-8F1EA329D420
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
*******************************************************************************
Forum Note: Use "Reply" to post a response in the discussion forum.
Bryce, did you check in the IIUG Yearly conferences presentations ? I am sure there should be a number of presentations about HDR and replication in general I'll post the URL ASAP, must check now