Re: Automated performance reporting on Linux on Z

2014-11-11 Thread Alan Altmark 
On Tuesday, 11/11/2014 at 01:44 EST, Alan Ackerman
 wrote:
> Would you be willing to tell us those "annoying idosyncrasies"?

No.  What annoys me may not annoy others, and in no case do the behaviors
diminish the value sufficiently to make me get rid of the product.  Like a
good TV program from my childhood, Good always triumphs over Evil.

> Our shop is very unlikely to move away from Velocity, but newer
customers, such
> as the Original Poster, might find it useful.

Since I spend most of my time in the System Management part of the world,
there are a lot of add-ons.  I recommend that my clients be savvy
shoppers, just as if they were in the market for a large-ticket item at
home.  To that end, always keep in mind:

1. A person's first love is often their best love.  It's not common sense,
but it's just the way humans work.  If you learn one of the programs
first, you will generally like it best since it's a huge improvement over
stone knives and bearskins.  But once learned, The Other Guy's warts look
worse in the shadow of your current selection.  Bottom line, when getting
comparisons from friends and family, make sure that they've actually spent
quality time with both.  ("I hate that car because Dad hates it" just
doesn't cut it.)

2. The vendor will tell you what's great about their product, and they
will says things repeatedly in the hope that if you hear it often enough
you will believe it to be True.  They will even provide reference
accounts.  See #1.  By the way, TALKING LOUDLY doesn't make a claim truer.

3. Beware of unsubstantiated claims, including references to "unbiased
opinion".  That phrase is an oxymoron.  Words like "most" and "everyone"
should raise a red flag.

4. Have actual requirements (gasp!).  Compare all claims against *your*
requirements, not those of the vendor.  If you don't care whether the car
is blue or not, ignore any color references from the manufacturer. Discuss
your requirements back at home before you go shopping so that your list is
complete and works for the whole family.  Get the big picture.

5. A product stands on its own merits.  If a salesperson tries to make
their case by telling what the other guy does wrong, they aren't saying
anything about their own product.

6. Products change.  Don't let ancient history or innuendo color your
decision-making process.  Claims from the vendor should be based on
reasonably current data.  ("JD Powers & Associates #1 Pick 3 Years on a
Row!" shouldn't have a footnote that says "1988-1990".)

7. You don't get what you pay for or deserve.  You get what you negotiate.
 Is there a trial period?  Is setup included?  Will they cart away your
old mattress?  For free?  Does it come in puce?  Are go-faster stripes
available?  But Caveat emptor.

8. Good products have a long list of unfulfilled requirements.  Bad
products have none because no one cares.  People who like the product the
best will have the most requirements as they are the ones who use it the
most.  So if someone complains about a product, dig deeper.

9. To Mr. Ackerman's point, give yourself a break and don't be afraid to
change your mind.  "Try it, you'll like it!"  Maybe not so much, eh?  Just
set Management's expectations properly ahead of time.

Feel free to also use the above for buying insurance, a car, or a set of
steak knives.

Alan Altmark

Senior Managing z/VM and Linux Consultant
Lab Services System z Delivery Practice
IBM Systems & Technology Group
ibm.com/systems/services/labservices
office: 607.429.3323
mobile; 607.321.7556
alan_altm...@us.ibm.com
IBM Endicott

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Re: Automated performance reporting on Linux on Z

2014-11-11 Thread Alan Ackerman 
Would you be willing to tell us those "annoying idosyncrasies"?

Our shop is very unlikely to move away from Velocity, but newer customers, such 
as the Original Poster, might find it useful.

> To successfully grow a zLinux farm, you *need* a commercial performance
> management solution for z/VM.   All of the solutions mentioned here are
> excellent.   Each has its own annoying idiosyncrasies, of course, but so
> does any other applications you buy.
> 
> Alan Altmark
> 
> Senior Managing z/VM and Linux Consultant
> Lab Services System z Delivery Practice
> IBM Systems & Technology Group
> ibm.com/systems/services/labservices
> office: 607.429.3323
> mobile; 607.321.7556
> alan_altm...@us.ibm.com
> IBM Endicott

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Re: Automated performance reporting on Linux on Z

2014-11-10 Thread Alan Altmark 
On Monday, 11/10/2014 at 04:40 EST, "Veencamp, Jonathon D."
 wrote:

> - "vmcp ind" command.  I gather AVGPROC for IFL hardware utilization.
And also
> XSTORE and PAGING info.
> - I also iteratively issue 'vmcp ind linuxguestname' for each of our
Linii, so
> we can display how much actual CPU each of the hosts is using.  On a low
> priority host, top and sar can give an incomplete picture if the
hipervisor is
> choking them off.

I always tell clients that CP INDICATE doesn't scale.   While those values
are fine when the system is humming along nicely (i.e. you don't really
care about them), they are generally of minimal use  when (not if) you
have trouble.  While they can tell you that users are waiting for I/O, for
example, they can't really give you insight into why.  And the bigger your
system, the less meaningful INDICATE becomes.

To successfully grow a zLinux farm, you *need* a commercial performance
management solution for z/VM.   All of the solutions mentioned here are
excellent.   Each has its own annoying idiosyncrasies, of course, but so
does any other applications you buy.

Alan Altmark

Senior Managing z/VM and Linux Consultant
Lab Services System z Delivery Practice
IBM Systems & Technology Group
ibm.com/systems/services/labservices
office: 607.429.3323
mobile; 607.321.7556
alan_altm...@us.ibm.com
IBM Endicott

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Re: Automated performance reporting on Linux on Z

2014-11-10 Thread Veencamp, Jonathon D. 
Mikael,

I also had to make a homemade monitor rather than spending money to get a real 
product, which I would have preferred.   In our experience, sar and the like 
are critically missing hypervisor stats that absolutely matter.  My duct tape 
and bailing wire solution was to configure one of our Linux guest with the 
security to run vmcp commands to display some basic ZVM stats, and I parse some 
of that info into the graphical monitor that displays all of this.

A couple suggestions/comments
- Linux under ZVM is nice that it reports CPU steal  (unlike Vmware).   And 
steal is pretty significant.
- "vmcp ind" command.  I gather AVGPROC for IFL hardware utilization.  And also 
XSTORE and PAGING info.
- I also iteratively issue 'vmcp ind linuxguestname' for each of our Linii, so 
we can display how much actual CPU each of the hosts is using.  On a low 
priority host, top and sar can give an incomplete picture if the hipervisor is 
choking them off.

Good luck!

Jon Veencamp



-Original Message-
From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Mikael 
Wargh
Sent: Monday, November 10, 2014 10:13 AM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: Automated performance reporting on Linux on Z

Thanks! This was very valuable information. I know that there are several 
commercial products for this purpose, but currently our calculations wont give 
us much more room to add additional licence expenses to this environment. 
Velocity software and IBM Tivoli tools are most probably the best available 
all-in-one products, but we just cannot afford that luxury yet. For z/VM 
monitoring and reporting we use z/VM's Performance reporter and we can also use 
our mainframe monitoring tools for this level of monitoring & automation. Most 
urgent need is however to provide Linux level reporting to our customers. For 
this purpose free nmon and sar tools seems to give sufficient information and 
they seems to be quite easy to implement. That's why the basic question was how 
to handle the last step of implementation of these tools aka automated web page 
or similar nmon/sar report generators which are available on AIX/Solaris and 
x86/x64 Linux worlds.

-Mikael

-Original Message-
From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Tito 
Garrido
Sent: 9. marraskuuta 2014 23:38
To: LINUX-390@VM.MARIST.EDU
Subject: Re: Automated performance reporting on Linux on Z

If you are going to try SAR take a look on KSAR project...
http://sourceforge.net/projects/ksar/

Regards,

Tito

On Sun, Nov 9, 2014 at 1:30 PM, Barton Robinson  wrote:

> I just had an installation demonstrate the perfect tool for you.  In
> looking at what alerts should be set for operations, one of their many
> zlinux servers had a swap full condition.  They were able to go back
> thru reports from last 12 months, took less than a minute to identify
> when the swap filled up, and what processes were running in the linux
> server at that time.  This was with a full web interface, all
> automated.  There were daily, weekly, monthly reports created for
> linux, z/vm, network all automatically and easily accessable thru their web 
> interface.
> If you would like to talk more, or maybe talk to some of the hundreds
> of installations that run this, that can be arranged.
>
>
>
> On 11/8/2014 2:36 PM, Mikael Wargh wrote:
>
>> Hello,
>>
>> We have several zLinuxes installed and been happy with the overall
>> functionality especially now when we got the new IBM wave tool installed.
>> However, Wave didn’t help is with the problem with the Linux level
>> reporting we currently have. z/VM performance reporter gives us
>> current overall status, but is not very helpful for longer period
>> capacity trending at least from single Linux perspective. Our
>> company’s standard Patrol agents show somewhat twisted and misleading
>> information about Linux on z especially on CPU point of view and this cannot 
>> be modified.
>>
>> So…now I’m trying to find out which Linux tool could be used as a
>> good base for this reporting requirement. Nmon seems to be the best
>> candidate so far, haven’t tested SAR yet. Currently I get automated
>> nmon reports from one of our zLinux servers and they generate nice
>> graphs with nmon_analyzer Excel program. However, it’s not very
>> convenient to manually create graphs every day on your
>> laptop…especially when we get reports from several servers in the future.
>>
>> On AIX you can use nmon2rrd which uses rrdtool for conversion. Also
>> there are some nmon2web scripts for AIX. For Linux you can get
>> several nmon web page generators which are often based on rrdtool and
>> you have also couple of viable SAR graph generating options. I have
>> tried to find a goo

Re: Automated performance reporting on Linux on Z

2014-11-10 Thread Mikael Wargh 
Thanks! This was very valuable information. I know that there are several 
commercial products for this purpose, but currently our calculations wont give 
us much more room to add additional licence expenses to this environment. 
Velocity software and IBM Tivoli tools are most probably the best available 
all-in-one products, but we just cannot afford that luxury yet. For z/VM 
monitoring and reporting we use z/VM's Performance reporter and we can also use 
our mainframe monitoring tools for this level of monitoring & automation. Most 
urgent need is however to provide Linux level reporting to our customers. For 
this purpose free nmon and sar tools seems to give sufficient information and 
they seems to be quite easy to implement. That's why the basic question was how 
to handle the last step of implementation of these tools aka automated web page 
or similar nmon/sar report generators which are available on AIX/Solaris and 
x86/x64 Linux worlds.

-Mikael

-Original Message-
From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Tito 
Garrido
Sent: 9. marraskuuta 2014 23:38
To: LINUX-390@VM.MARIST.EDU
Subject: Re: Automated performance reporting on Linux on Z

If you are going to try SAR take a look on KSAR project...
http://sourceforge.net/projects/ksar/

Regards,

Tito

On Sun, Nov 9, 2014 at 1:30 PM, Barton Robinson  wrote:

> I just had an installation demonstrate the perfect tool for you.  In 
> looking at what alerts should be set for operations, one of their many 
> zlinux servers had a swap full condition.  They were able to go back 
> thru reports from last 12 months, took less than a minute to identify 
> when the swap filled up, and what processes were running in the linux 
> server at that time.  This was with a full web interface, all 
> automated.  There were daily, weekly, monthly reports created for 
> linux, z/vm, network all automatically and easily accessable thru their web 
> interface.
> If you would like to talk more, or maybe talk to some of the hundreds 
> of installations that run this, that can be arranged.
>
>
>
> On 11/8/2014 2:36 PM, Mikael Wargh wrote:
>
>> Hello,
>>
>> We have several zLinuxes installed and been happy with the overall 
>> functionality especially now when we got the new IBM wave tool installed.
>> However, Wave didn’t help is with the problem with the Linux level 
>> reporting we currently have. z/VM performance reporter gives us 
>> current overall status, but is not very helpful for longer period 
>> capacity trending at least from single Linux perspective. Our 
>> company’s standard Patrol agents show somewhat twisted and misleading 
>> information about Linux on z especially on CPU point of view and this cannot 
>> be modified.
>>
>> So…now I’m trying to find out which Linux tool could be used as a 
>> good base for this reporting requirement. Nmon seems to be the best 
>> candidate so far, haven’t tested SAR yet. Currently I get automated 
>> nmon reports from one of our zLinux servers and they generate nice 
>> graphs with nmon_analyzer Excel program. However, it’s not very 
>> convenient to manually create graphs every day on your 
>> laptop…especially when we get reports from several servers in the future.
>>
>> On AIX you can use nmon2rrd which uses rrdtool for conversion. Also 
>> there are some nmon2web scripts for AIX. For Linux you can get 
>> several nmon web page generators which are often based on rrdtool and 
>> you have also couple of viable SAR graph generating options. I have 
>> tried to find a good solution for zLinux but so far haven’t been able 
>> to fill this automated web page generation gap.
>>
>> To sum this up:
>> - We need a performance reporting tool which can be fully automated 
>> via Linux scripting or similar for multiple zLinux instances. As 
>> there seems to be several freeware possibilities, we would prefer them.
>> - Currently I have created a routine which creates daily and weekly 
>> nmon scripts which are then collected to central zLinux reporting server.
>> - Those nmon files are useful to our maintenance people, but we need 
>> also simple graphs to other parties preferably in some picture/html format.
>> - Ideal solution would be some simple way to convert those nmon files 
>> to html pages automatically as nmon2rrd does.
>> - One possibility would be to upload those nmon files to some AIX/x64 
>> server and do the conversion there, but it would create extra data 
>> transfer and to be honest would be quite embarrassing from zLinux point of 
>> view.
>> - Is there any alternative routes to do this or should I try to 
>> compile some of these AIX/x64 tools to syst

Re: Automated performance reporting on Linux on Z

2014-11-10 Thread Bill Bitner 
The IBM OMEGAMON XE on z/VM and Linux tool collects performance metrics and
allows customers to store them in a data warehouse (included).  Customers
choose what to keep and they summarize and prune the data based on the
reporting they would like to consider.  Another included feature of any
OMEGAMON or IBM Tivoli Monitoring component is the Cognos based reporting
tool Tivoli Common Reporter (TCR) which enables custom and dynamic reporting
against the data in the data warehouse.  Many customers use this to produce
the kind of reports you are mentioning.  Not only can you report on z/VM and
Linux data, but any other data in the data warehouse.  If you collect middleware
or database data on those Linux guests, these attributes can be included in
this reporting structure.  Finally, another included feature that uses this
same warehouse data is the Tivoli Performance Analyzer which adds vendor
supplied predictive capability focusing on future performance issues.  Customers
can change values and play "what if" type of scenarios.

All in all, you get real-time performance analysis, custom reporting, and
predictive analysis in one tool.


Bill Bitner

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Re: Automated performance reporting on Linux on Z

2014-11-10 Thread Andrew Porter 
I've had customers for the product I work on complain of performance issues and 
a number of times when the Linux server was examined the source of the problem 
was a commercial monitoring agent that had been in a run loop for weeks. A Java 
process so not quite as light-weight as you would expect when you hear the word 
"agent". Start small and use an extended trial to be sure that your attempt at 
a cure does not make the disease worse...


Andrew




 From: Tito Garrido 
To: LINUX-390@VM.MARIST.EDU 
Sent: Sunday, November 9, 2014 1:38 PM
Subject: Re: Automated performance reporting on Linux on Z
 

If you are going to try SAR take a look on KSAR project...
http://sourceforge.net/projects/ksar/

Regards,

Tito

On Sun, Nov 9, 2014 at 1:30 PM, Barton Robinson  wrote:

> I just had an installation demonstrate the perfect tool for you.  In
> looking at what alerts should be set for operations, one of their many
> zlinux servers had a swap full condition.  They were able to go back thru
> reports from last 12 months, took less than a minute to identify when the
> swap filled up, and what processes were running in the linux server at that
> time.  This was with a full web interface, all automated.  There were
> daily, weekly, monthly reports created for linux, z/vm, network all
> automatically and easily accessable thru their web interface.
> If you would like to talk more, or maybe talk to some of the hundreds of
> installations that run this, that can be arranged.
>
>
>
> On 11/8/2014 2:36 PM, Mikael Wargh wrote:
>
>> Hello,
>>
>> We have several zLinuxes installed and been happy with the overall
>> functionality especially now when we got the new IBM wave tool installed.
>> However, Wave didn’t help is with the problem with the Linux level
>> reporting we currently have. z/VM performance reporter gives us current
>> overall status, but is not very helpful for longer period capacity trending
>> at least from single Linux perspective. Our company’s standard Patrol
>> agents show somewhat twisted and misleading information about Linux on z
>> especially on CPU point of view and this cannot be modified.
>>
>> So…now I’m trying to find out which Linux tool could be used as a good
>> base for this reporting requirement. Nmon seems to be the best candidate so
>> far, haven’t tested SAR yet. Currently I get automated nmon reports from
>> one of our zLinux servers and they generate nice graphs with nmon_analyzer
>> Excel program. However, it’s not very convenient to manually create graphs
>> every day on your laptop…especially when we get reports from several
>> servers in the future.
>>
>> On AIX you can use nmon2rrd which uses rrdtool for conversion. Also there
>> are some nmon2web scripts for AIX. For Linux you can get several nmon web
>> page generators which are often based on rrdtool and you have also couple
>> of viable SAR graph generating options. I have tried to find a good
>> solution for zLinux but so far haven’t been able to fill this automated web
>> page generation gap.
>>
>> To sum this up:
>> - We need a performance reporting tool which can be fully automated via
>> Linux scripting or similar for multiple zLinux instances. As there seems to
>> be several freeware possibilities, we would prefer them.
>> - Currently I have created a routine which creates daily and weekly nmon
>> scripts which are then collected to central zLinux reporting server.
>> - Those nmon files are useful to our maintenance people, but we need also
>> simple graphs to other parties preferably in some picture/html format.
>> - Ideal solution would be some simple way to convert those nmon files to
>> html pages automatically as nmon2rrd does.
>> - One possibility would be to upload those nmon files to some AIX/x64
>> server and do the conversion there, but it would create extra data transfer
>> and to be honest would be quite embarrassing from zLinux point of view.
>> - Is there any alternative routes to do this or should I try to compile
>> some of these AIX/x64 tools to system z Linux?
>> - We use Redhat 6.4 and z/VM 6.3
>>
>> Best regards,
>> Mikael Wargh
>>
>
> --
> For LINUX-390 subscribe / signoff / archive access instructions,
> send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or
> visit
> http://www.marist.edu/htbin/wlvindex?LINUX-390
> --
> For more information on Linux on System z, visit
> http://wiki.linuxvm.org/
>

Re: Automated performance reporting on Linux on Z

2014-11-10 Thread Tracy Dean 
The IBM OMEGAMON XE on z/VM and Linux tool collects performance metrics
and allows customers to store them in a data warehouse (included).
Customers choose what to keep and they summarize and prune the data based
on the reporting they would like to consider. Another included feature of
any OMEGAMON or IBM Tivoli Monitoring component is the Cognos based
reporting tool Tivoli Common Reporter (TCR) which enables custom and
dynamic reporting against the data in the data warehouse. Many customers
use this to produce the kind of reports you are mentioning. Not only can
you report on z/VM and Linux data, but any other data in the data
warehouse. If you collect middleware or database data on those Linux
guests, these attributes can be included in this reporting structure.
Finally, another included feature that uses this same warehouse data is
the Tivoli Performance Analyzer which adds vendor supplied predictive
capability focusing on future performance issues. Customers can change
values and play "what if" type of scenarios. All in all, you get real-time
performance analysis, custom reporting, and predictive analysis in one
tool set.

Tracy Dean, IBM

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Re: Automated performance reporting on Linux on Z

2014-11-09 Thread Barton Robinson 
The problem with these tools is other than the pretty charts, do they 
help solve the problems?  That ability to solve problems comes with 
experience - difficult to write tools to solve problems if having no 
experience with those problems.  Installations that run Linux on 'Z' 
have greater requirements than distributed servers - running at higher 
utilization, expected for Z, requires complete data, trend data, 
automatic data capture.  The platform has been around for near 15 years 
- with lots of experience available.  Can you go back months with a tool 
and see when the problem started, and what caused it?  Or do you just 
reboot?


On 11/9/2014 1:38 PM, Tito Garrido wrote:

If you are going to try SAR take a look on KSAR project...
http://sourceforge.net/projects/ksar/

Regards,

Tito

On Sun, Nov 9, 2014 at 1:30 PM, Barton Robinson 
wrote:
I just had an installation demonstrate the perfect tool for you.  In
looking at what alerts should be set for operations, one of their many
zlinux servers had a swap full condition.  They were able to go back thru
reports from last 12 months, took less than a minute to identify when the
swap filled up, and what processes were running in the linux server at that
time.  This was with a full web interface, all automated.  There were
daily, weekly, monthly reports created for linux, z/vm, network all
automatically and easily accessable thru their web interface.
If you would like to talk more, or maybe talk to some of the hundreds of
installations that run this, that can be arranged.



On 11/8/2014 2:36 PM, Mikael Wargh wrote:


Hello,

We have several zLinuxes installed and been happy with the overall
functionality especially now when we got the new IBM wave tool installed.
However, Wave didn’t help is with the problem with the Linux level
reporting we currently have. z/VM performance reporter gives us current
overall status, but is not very helpful for longer period capacity trending
at least from single Linux perspective. Our company’s standard Patrol
agents show somewhat twisted and misleading information about Linux on z
especially on CPU point of view and this cannot be modified.

So…now I’m trying to find out which Linux tool could be used as a good
base for this reporting requirement. Nmon seems to be the best candidate so
far, haven’t tested SAR yet. Currently I get automated nmon reports from
one of our zLinux servers and they generate nice graphs with nmon_analyzer
Excel program. However, it’s not very convenient to manually create graphs
every day on your laptop…especially when we get reports from several
servers in the future.

On AIX you can use nmon2rrd which uses rrdtool for conversion. Also there
are some nmon2web scripts for AIX. For Linux you can get several nmon web
page generators which are often based on rrdtool and you have also couple
of viable SAR graph generating options. I have tried to find a good
solution for zLinux but so far haven’t been able to fill this automated web
page generation gap.

To sum this up:
- We need a performance reporting tool which can be fully automated via
Linux scripting or similar for multiple zLinux instances. As there seems to
be several freeware possibilities, we would prefer them.
- Currently I have created a routine which creates daily and weekly nmon
scripts which are then collected to central zLinux reporting server.
- Those nmon files are useful to our maintenance people, but we need also
simple graphs to other parties preferably in some picture/html format.
- Ideal solution would be some simple way to convert those nmon files to
html pages automatically as nmon2rrd does.
- One possibility would be to upload those nmon files to some AIX/x64
server and do the conversion there, but it would create extra data transfer
and to be honest would be quite embarrassing from zLinux point of view.
- Is there any alternative routes to do this or should I try to compile
some of these AIX/x64 tools to system z Linux?
- We use Redhat 6.4 and z/VM 6.3

Best regards,
Mikael Wargh


--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or
visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/






--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Re: Automated performance reporting on Linux on Z

2014-11-09 Thread Tito Garrido 
If you are going to try SAR take a look on KSAR project...
http://sourceforge.net/projects/ksar/

Regards,

Tito

On Sun, Nov 9, 2014 at 1:30 PM, Barton Robinson  wrote:

> I just had an installation demonstrate the perfect tool for you.  In
> looking at what alerts should be set for operations, one of their many
> zlinux servers had a swap full condition.  They were able to go back thru
> reports from last 12 months, took less than a minute to identify when the
> swap filled up, and what processes were running in the linux server at that
> time.  This was with a full web interface, all automated.  There were
> daily, weekly, monthly reports created for linux, z/vm, network all
> automatically and easily accessable thru their web interface.
> If you would like to talk more, or maybe talk to some of the hundreds of
> installations that run this, that can be arranged.
>
>
>
> On 11/8/2014 2:36 PM, Mikael Wargh wrote:
>
>> Hello,
>>
>> We have several zLinuxes installed and been happy with the overall
>> functionality especially now when we got the new IBM wave tool installed.
>> However, Wave didn’t help is with the problem with the Linux level
>> reporting we currently have. z/VM performance reporter gives us current
>> overall status, but is not very helpful for longer period capacity trending
>> at least from single Linux perspective. Our company’s standard Patrol
>> agents show somewhat twisted and misleading information about Linux on z
>> especially on CPU point of view and this cannot be modified.
>>
>> So…now I’m trying to find out which Linux tool could be used as a good
>> base for this reporting requirement. Nmon seems to be the best candidate so
>> far, haven’t tested SAR yet. Currently I get automated nmon reports from
>> one of our zLinux servers and they generate nice graphs with nmon_analyzer
>> Excel program. However, it’s not very convenient to manually create graphs
>> every day on your laptop…especially when we get reports from several
>> servers in the future.
>>
>> On AIX you can use nmon2rrd which uses rrdtool for conversion. Also there
>> are some nmon2web scripts for AIX. For Linux you can get several nmon web
>> page generators which are often based on rrdtool and you have also couple
>> of viable SAR graph generating options. I have tried to find a good
>> solution for zLinux but so far haven’t been able to fill this automated web
>> page generation gap.
>>
>> To sum this up:
>> - We need a performance reporting tool which can be fully automated via
>> Linux scripting or similar for multiple zLinux instances. As there seems to
>> be several freeware possibilities, we would prefer them.
>> - Currently I have created a routine which creates daily and weekly nmon
>> scripts which are then collected to central zLinux reporting server.
>> - Those nmon files are useful to our maintenance people, but we need also
>> simple graphs to other parties preferably in some picture/html format.
>> - Ideal solution would be some simple way to convert those nmon files to
>> html pages automatically as nmon2rrd does.
>> - One possibility would be to upload those nmon files to some AIX/x64
>> server and do the conversion there, but it would create extra data transfer
>> and to be honest would be quite embarrassing from zLinux point of view.
>> - Is there any alternative routes to do this or should I try to compile
>> some of these AIX/x64 tools to system z Linux?
>> - We use Redhat 6.4 and z/VM 6.3
>>
>> Best regards,
>> Mikael Wargh
>>
>
> --
> For LINUX-390 subscribe / signoff / archive access instructions,
> send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or
> visit
> http://www.marist.edu/htbin/wlvindex?LINUX-390
> --
> For more information on Linux on System z, visit
> http://wiki.linuxvm.org/
>



-- 

Linux User #387870
.
 _/_õ|__|
..º[ .-.___.-._| . . . .
.__( o)__( o).:___

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Re: Automated performance reporting on Linux on Z

2014-11-09 Thread Barton Robinson 
I just had an installation demonstrate the perfect tool for you.  In 
looking at what alerts should be set for operations, one of their many 
zlinux servers had a swap full condition.  They were able to go back 
thru reports from last 12 months, took less than a minute to identify 
when the swap filled up, and what processes were running in the linux 
server at that time.  This was with a full web interface, all 
automated.  There were daily, weekly, monthly reports created for linux, 
z/vm, network all automatically and easily accessable thru their web 
interface.
If you would like to talk more, or maybe talk to some of the hundreds of 
installations that run this, that can be arranged.




On 11/8/2014 2:36 PM, Mikael Wargh wrote:

Hello,

We have several zLinuxes installed and been happy with the overall 
functionality especially now when we got the new IBM wave tool installed. 
However, Wave didn’t help is with the problem with the Linux level reporting we 
currently have. z/VM performance reporter gives us current overall status, but 
is not very helpful for longer period capacity trending at least from single 
Linux perspective. Our company’s standard Patrol agents show somewhat twisted 
and misleading information about Linux on z especially on CPU point of view and 
this cannot be modified.

So…now I’m trying to find out which Linux tool could be used as a good base for 
this reporting requirement. Nmon seems to be the best candidate so far, haven’t 
tested SAR yet. Currently I get automated nmon reports from one of our zLinux 
servers and they generate nice graphs with nmon_analyzer Excel program. 
However, it’s not very convenient to manually create graphs every day on your 
laptop…especially when we get reports from several servers in the future.

On AIX you can use nmon2rrd which uses rrdtool for conversion. Also there are 
some nmon2web scripts for AIX. For Linux you can get several nmon web page 
generators which are often based on rrdtool and you have also couple of viable 
SAR graph generating options. I have tried to find a good solution for zLinux 
but so far haven’t been able to fill this automated web page generation gap.

To sum this up:
- We need a performance reporting tool which can be fully automated via Linux 
scripting or similar for multiple zLinux instances. As there seems to be 
several freeware possibilities, we would prefer them.
- Currently I have created a routine which creates daily and weekly nmon 
scripts which are then collected to central zLinux reporting server.
- Those nmon files are useful to our maintenance people, but we need also 
simple graphs to other parties preferably in some picture/html format.
- Ideal solution would be some simple way to convert those nmon files to html 
pages automatically as nmon2rrd does.
- One possibility would be to upload those nmon files to some AIX/x64 server 
and do the conversion there, but it would create extra data transfer and to be 
honest would be quite embarrassing from zLinux point of view.
- Is there any alternative routes to do this or should I try to compile some of 
these AIX/x64 tools to system z Linux?
- We use Redhat 6.4 and z/VM 6.3

Best regards,
Mikael Wargh


--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Re: Automated performance reporting on Linux on Z

2014-11-08 Thread Rick Troth 
On 11/08/2014 05:36 PM, Mikael Wargh wrote:
> - We need a performance reporting tool which can be fully automated via Linux 
> scripting or similar for multiple zLinux instances. As there seems to be 
> several freeware possibilities, we would prefer them.
> - Currently I have created a routine which creates daily and weekly nmon 
> scripts which are then collected to central zLinux reporting server.
> - Those nmon files are useful to our maintenance people, but we need also 
> simple graphs to other parties preferably in some picture/html format.
> - Ideal solution would be some simple way to convert those nmon files to html 
> pages automatically as nmon2rrd does.
> - One possibility would be to upload those nmon files to some AIX/x64 server 
> and do the conversion there, but it would create extra data transfer and to 
> be honest would be quite embarrassing from zLinux point of view.
> - Is there any alternative routes to do this or should I try to compile some 
> of these AIX/x64 tools to system z Linux?
> - We use Redhat 6.4 and z/VM 6.3

Velocity's zVPS handles performance monitoring and alerting for both the
Linux side and the z/VM side. It conveys Linux metrics via SNMP into the
common monitor stream on the z/VM host, and correlates the hypervisor's
view of guest performance.

Check out the suite of tools at ...

http://www.velocitysoftware.com/product.html




--

Rick Troth
Senior Software Developer

Velocity Software Inc.
Mountain View, CA 94041
Main: (877) 964-8867
Direct: (614) 594-9768
ri...@velocitysoftware.com 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
For more information on Linux on System z, visit
http://wiki.linuxvm.org/