unsubscribe

2020-06-17 Thread Brown, Duncan
Unsubscribe 

Disclaimer

The information contained in this communication from the sender is 
confidential. It is intended solely for use by the recipient and others 
authorized to receive it. If you are not the recipient, you are hereby notified 
that any disclosure, copying, distribution or taking action in relation of the 
contents of this information is strictly prohibited and may be unlawful.

This email has been scanned for viruses and malware, and may have been 
automatically archived by Mimecast Ltd, an innovator in Software as a Service 
(SaaS) for business. Providing a safer and more useful place for your human 
generated data. Specializing in; Security, archiving and compliance. To find 
out more visit the Mimecast website.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www2.marist.edu/htbin/wlvindex?LINUX-390


Re: Running multiple Oracle on a single Linux guest

2020-06-17 Thread Herald ten Dam
We have the setup of 1 oracle database on 1 linux server. This way we only plan 
to use the memory just what Oracle needs (and some for Linux and monitoring 
stuff). We have always 2 vdisks for swap and then 1 mdisk as swap.
For your case I think you can have memory at the Oracle requirement (SGA+PGA) 
and give a swap on vdisk for some amount.

Also depends on version, if you have Oracle 12 or above and use Multitenant 
then SGA+PGA is more shared among the databases. Normally by using multitenant 
memory can be less than if used standalone. Multitenant is free till 3 
pluggable databases, so If you have more, more instances are still needed (or 
pay for it).

Met vriendelijke groet,
Herald ten Dam
Database Specialist Oracle, Medior z/VM

ICU IT Services BV
Transistorstraat 55b  I  1322 CK  ALMERE

M 06 – 28 891 653  I  E 
herald.ten@icu-it.nl
T 088 – 5 234 123  I  
www.icu-it.nl
  I  KvK 32135776


Van: Linux on 390 Port  namens Rob van der Heij 

Verzonden: woensdag 17 juni 2020 13:29
Aan: LINUX-390@VM.MARIST.EDU 
Onderwerp: Re: Running multiple Oracle on a single Linux guest

On Wed, 17 Jun 2020 at 09:14, Peter  wrote:

> So I can't cap memory based on a oracle instance ? Like prioritising each
> oracle instance ?
>

You can to some extent. You define the maximum by SGA and PGA per instance,
and things like connection pools. The actual usage varies somewhat by
workload. But very few customers will make their guests large enough to
hold all database instances at their maximum usage. That would leave a lot
of excess memory that Linux will use for other things (like cache data).
Instead, they share much of the white space and cross their fingers.

If your databases have short burst of activity and long periods of idle, it
might be more attractive to configure Linux guests to fit that instance and
let z/VM take resources away when there is demand from others and the
instance is idle.

Rob

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www2.marist.edu/htbin/wlvindex?LINUX-390

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www2.marist.edu/htbin/wlvindex?LINUX-390


Re: Running multiple Oracle on a single Linux guest

2020-06-17 Thread Rob van der Heij
On Wed, 17 Jun 2020 at 09:14, Peter  wrote:

> So I can't cap memory based on a oracle instance ? Like prioritising each
> oracle instance ?
>

You can to some extent. You define the maximum by SGA and PGA per instance,
and things like connection pools. The actual usage varies somewhat by
workload. But very few customers will make their guests large enough to
hold all database instances at their maximum usage. That would leave a lot
of excess memory that Linux will use for other things (like cache data).
Instead, they share much of the white space and cross their fingers.

If your databases have short burst of activity and long periods of idle, it
might be more attractive to configure Linux guests to fit that instance and
let z/VM take resources away when there is demand from others and the
instance is idle.

Rob

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www2.marist.edu/htbin/wlvindex?LINUX-390


Re: Running multiple Oracle on a single Linux guest

2020-06-17 Thread Peter
So I can't cap memory based on a oracle instance ? Like prioritising each
oracle instance ?



On Wed, 17 Jun, 2020, 10:44 AM Rob van der Heij,  wrote:

> On Wed, 17 Jun 2020 at 07:06, Peter  wrote:
>
> We have 2 IFLs for zVM(non SSI). We are planning to run multiple instances
> > of oracle on a single Linux guest server.
> >
> > Almost these oracle are expected to have a very low workload.
> >
> > Is anyone in the group who are running a similar workload and is there
> any
> > bottle neck to check while running multiple oracle in a single linux
> guest.
> >
>
> Yes, that is a typical situation because sometimes during a project people
> come up with more database instances they need, and don't want to go back
> and get additional servers. Since the database instances are low usage, it
> is tempting to believe you can share the white space in the server.
> However, sometimes your database server will peak and show high usage, and
> when several instances get busy at the same time you may run out of memory.
> You probably should not configure PGA and SGA for each instance as if it
> were the only instance in that virtual machine. Make sure you configure the
> database not to absorb all resources (like connection pool size, SGA and
> PGA, etc).
>
> There's a baseline cost per virtual machine in resources (mostly memory)
> and management (deployment, maintenance, monitoring). When your database
> instances are really small with tiny SGA and PGA and few connections, that
> baseline cost can get proportionally high. When you have advanced tools to
> manage the guests, it is easier to decide to split the workload into
> different machines. On the other hand, if you mix different workloads in
> the same Linux guest, you can't use z/VM level tools and settings to
> prioritize a specific workload or do charge back on that level.
>
> Rob
>
> --
> For LINUX-390 subscribe / signoff / archive access instructions,
> send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or
> visit
> http://www2.marist.edu/htbin/wlvindex?LINUX-390
>

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www2.marist.edu/htbin/wlvindex?LINUX-390


Re: Running multiple Oracle on a single Linux guest

2020-06-17 Thread Rob van der Heij
On Wed, 17 Jun 2020 at 07:06, Peter  wrote:

We have 2 IFLs for zVM(non SSI). We are planning to run multiple instances
> of oracle on a single Linux guest server.
>
> Almost these oracle are expected to have a very low workload.
>
> Is anyone in the group who are running a similar workload and is there any
> bottle neck to check while running multiple oracle in a single linux guest.
>

Yes, that is a typical situation because sometimes during a project people
come up with more database instances they need, and don't want to go back
and get additional servers. Since the database instances are low usage, it
is tempting to believe you can share the white space in the server.
However, sometimes your database server will peak and show high usage, and
when several instances get busy at the same time you may run out of memory.
You probably should not configure PGA and SGA for each instance as if it
were the only instance in that virtual machine. Make sure you configure the
database not to absorb all resources (like connection pool size, SGA and
PGA, etc).

There's a baseline cost per virtual machine in resources (mostly memory)
and management (deployment, maintenance, monitoring). When your database
instances are really small with tiny SGA and PGA and few connections, that
baseline cost can get proportionally high. When you have advanced tools to
manage the guests, it is easier to decide to split the workload into
different machines. On the other hand, if you mix different workloads in
the same Linux guest, you can't use z/VM level tools and settings to
prioritize a specific workload or do charge back on that level.

Rob

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www2.marist.edu/htbin/wlvindex?LINUX-390