Scaling of VM Roles and Worker Roles, why divide by total?

I am trying to tweak with the auto-scale settings and it is my understanding that the CPU load is divided among the total number of instances within the Role.
I get this in concept.
However, what I am seeing is that if I have two instances in my availability group, and I have one instance running. That if I set the CPU load slider too high, my new instance never turns on / provisions.
It is as if the calculation is always averaging among the total possible - therefore if one is at 100% CPU, then that equates to 50% CPU. And if my slider is set to 60%, my second VM never turns on. (I left it running all night this way once to see).
Where what I think the slider should mean is that I never want a single instance to go beyond 60% utilization, if it does then give me another instance.
Am I following this properly?
Since I only have the scaling metric of CPU (since I don't have a Service Bus queue) I have to use it wisely and make sure that it reacts when appropriate - and in my case I might want it to scale out at 40% to ensure that there is no negative
user experience.
And right now, I am not seeing that I can make that happen.
Brian Ehlert<br/> http://ITProctology.blogspot.com<br/> Learn. Apply. Repeat. <br/&gt

hi
autoscaling per CPU is compared with the average CPU usage of all instances of the role.
so if you set the bottom bar is 20% and have 2 instances running there with CPU usage 37% and 4%, azure fabrci will autoscale it to 3 instances, because average cpu is 20.5% and beyond the bottom bar.
actually, autoscaling based on cpu works as below:
All instances are included when calculating the average percentage of CPU usage and the
average is based on use over the previous hour. Depending on the number of instances that your application is using, it can take longer than the specified wait time for the scale action to occur if the wait time is set
very low.
As a result, if you have an app that is at 0% load, and then start running a load test to make it go to 100% (and have a scale-up target of 80%) it will take at least 45 minutes before
the scale action will start. However, this is not a typical real-world scenario. It’s more likely that your load is already high (say 75%). In this scenario, it would take much less time to trigger the scale action.
One of the reasons that we do an hourly average is, with the current platform, it’s impossible to get metrics from Virtual Machines or Cloud Services under a 15 minute latency.
So, if we scaled based just on the last 5 or 10 minutes, we would never have data to scale on. You can see this in the screenshot below, it was taken around 5:30, but the most recent data point is at 5:15.
In the future, the azure platform is looking at ways to speed up metric collection, but this is likely not coming for quite some time. As a result, the best we can do is a rolling
average over a larger time window.
let me know if there is any question.
best regards
Jian

Similar Messages

Share Same DbContext between web role and worker role

HI
My "Windows Azure Cloud Service" project consists of one web role (.NET MVC) and one worker role, they are working fine under "Emulator".
The web role project is orginally a Azure web site application which is converted to a web role project and the connectionstring is located in the web role application
http://thinkfirstcodelater.com/blog/?p=1922
When I try to access the dbcontext in worker role like this
public override void Run()
            // This is a sample worker implementation. Replace with your logic.
            Trace.TraceInformation("CCDWorker entry point called", "Information");
            while (true)
                Thread.Sleep(50000);
MyContext context = new MyContext();
                List<MyEntity> ccds = context.MyEntity.ToList();
                Trace.TraceInformation("Working", "Information");
I got this exception: "The model backing the context has changed since the database was created." Please help
Regards
Mark

Hi Mark,
Thanks for posting!
Like jeff said in this blog(http://weblogs.asp.net/scottgu/archive/2010/08/03/using-ef-code-first-with-an-existing-database.aspx ),When a model
is first created, we run a DatabaseInitializer to do things like create the database if it's not there or add seed data. The default DatabaseInitializer tries to compare the database schema needed to use the model with a hash of the schema stored in an EdmMetadata
table that is created with a database (when Code First is the one creating the database). Existing databases won’t have the EdmMetadata table and so won’t have the hash…and the implementation today will throw if that table is missing. We'll work on changing
this behavior before we ship the fial version since it is the default. Until then, existing databases do not generally need any database initializer so it can be turned off for your context type by calling:
Database.SetInitializer<YourDatabaseContext>(null);
Also, you could refer to this thread,http://stackoverflow.com/questions/3600175/the-model-backing-the-database-context-has-changed-since-the-database-was-crea
Regards,
We are trying to better understand customer views on social support experience, so your participation in this interview project would be greatly appreciated if you have time. Thanks for helping make community forums a great place.
Click
HERE to participate the survey.

August 2014 Scheduled Maintenance - Are Web and Worker Roles affected?

Are Web and Worker Roles affected by the Scheduled Maintenance that will be performed on Aug 2, 2014? The email that I received from Microsoft Azure states that it will affect VMs and Cloud Services and goes on to say how to keep your VMs available
despite the maintenance by setting up Availability Sets for VMs, but does not mention a strategy for Web and Worker roles.
What am I missing here?

Hello Farid_Azure,
Yes, the Web and Worker Roles are also affected by the scheduled maintenance as the web roles and worker roles are residing on the hardware that will go on maintenance.
The Cloud Service deployments will also reboot during the maintenance. You can increase the number of instances running for your web roles and worker roles to avoid being affected:
http://azure.microsoft.com/en-us/documentation/articles/cloud-services-how-to-scale/#manualscale
I hope that this answers your question.
Thanks,
Syed Irfan Hussain

What are Azure limitations for Websockets in Cloud Services (web and worker role)?

A WebSocket Server should be built on Azure platform with OnPrem connections and have questions regarding limitations for Websockets in Azure Cloud Services - web and worker roles.
Websockets can be configured for Web Sites and limitations are understood, but Azure Websites is not an option.
Nevertheless it is planned to run a web service (without UI - no web site) as a Cloud service which has secure websocket (WSS) connections to OnPrem machines. Websocket protocol is enabled for IIS8 on Cloud services web and worker roles. Azure Service Bus Relay
is not an option.
Questions:
1) Are Websockets supported for Azure Cloud services web and worker roles? we assume yes
2) What are potential limitations from Azure side to support concurrent Websocket connections? We are aware that CPU, memory etc are limitations, but are there additional limitations from MS Azure side?

Hi,
As I know, azure cloud service web and worker role support Websockets, users can connect to the role via the special endpoint, if we use Azure cloud service, I think we can monitor the metrics such as CPU, memory, etc... and scale our cloud service via these
metrics to keep the websockets working, refer to
http://azure.microsoft.com/en-us/documentation/articles/cloud-services-how-to-scale/ for more information about how to scale a cloud service.
Regards

Security-role and security-role-assignment not working in WL7.0

Hello all..
Some EJB components that worked fine in WebLogic 6.1 no longer work in
WL7.0. It has to do with the security-role and security-role-assignment
descriptor elements no longer allowing anonymous users to be included in the
authorization for a bean.
For example, in WL6.1 placing these items in ejb-jar.xml:
<assembly-descriptor>
<security-role>
<role-name>Employees</role-name>
</security-role>
<method-permission>
<role-name>Employees</role-name>
<method>
<ejb-name>CustomerEJB</ejb-name>
<method-name>*</method-name>
</method>
</method-permission>
and mapping WebLogic default users to this role in weblogic-ejb-jar.xml:
<security-role-assignment>
<role-name>Employees</role-name>
<principal-name>guest</principal-name>
<principal-name>system</principal-name>
</security-role-assignment>
worked fine for clients creating their context using a simple
InitialContext() constructor without specifying SECURITY_PRINCIPAL or
SECURITY_CREDENTIALS. These users were basically "guest" to WebLogic, and
the security-role-assignment element above told WebLogic that "guest" was in
the Employees role for purposes of this EJB archive.
Worked in WL6.1, no longer works in WL7.0. Client receives typical
permission exception:
java.rmi.AccessException: Security violation: insufficient permission to
access method 'create'
If I explicity connect as "system" things are fine, or I can create a new
user in the default realm in WebLogic, put a matching <principal-name>
element in the section above, and connect as that user. Note that if I leave
off the <security-role> section completely, or set the required role name to
"everyone", the anonymous access works fine. Apparently the anonymous user
is a member of "everyone" behind the scenes even though "everyone" does not
appear in the realm list of groups or roles.
So, my question boils down to this: Is there a "magic" username in WL7 like
"guest" was in WL6.1 that can be mapped to the required role name, or must
every client connection use a true weblogic-created user with appropriate
role assignments used to map it to the required role name.
-Greg
P.S. Note that none of the EJB examples provided with WL used
<security-role>..
Check out my WebLogic 6.1 Workbook for O'Reilly EJB Third Edition
www.amazon.com/exec/obidos/ASIN/1931822468 or www.titan-books.com

Below are the screen shots for PFCG:

When to use "my role" and "partner role" in BPEL?

I'm a bit confused when to set/use partner role and my role.
Can anyone shed a little light,
regards, Henrik

Saurabh,
> I humbly disagree with your explanation of inputs
No need to be humble, you can boldy disagree. :)
You're right that I did technically use the wrong term in that sentence of my explanation. I updated the post and corrected it. However the gist of what I was saying is still true.
There are two invocation types. People use different terms but here I'll call them request-response and one-way. A request-response invocation type is used for what we typically think of as a "synchronous" process. That is, the service consumer is blocked until the service responds. It's like methodA() in Java calling methodB(). methodA() is blocked until methodB() completes. (In fact, this is exactly what it's like since all invocations on our BPEL engine ultimately go through our Java API.)
In the case of one-way, the service consumer is not blocked. This is often referred to as fire-and-forget. It simply sends its request, then it is free to continue or do whatever it wants. Moreover, nothing is returned to the client (you fired-and-forgot, remember). Typical "asynchronous" BPEL processes uses this invocation type.
So we have those two invocation types. Yet the problem before us how to have an asynchronous process return a result. You can't use request-response because the service consumer is blocked until the process finishes. You can't just use a one-way because nothing is returned to the caller. What to do?
The way the BPEL standard solved the problem is to use two one-way invocation types. The first one is to invoke the process. The second one is a one-way from the BPEL process to the service consumer to return the result. There are some glaring implications of this:
1. When the BPEL process returns it's result, things have now switched: The BPEL process becomes the service consumer, and the (original) service consumer becomes the service.
2. The service consumer has to be able to listen for one-way invocation type requests.
3. The BPEL process has to know how and where to call the service consumer back. This information is passed in the original request. As well as containing the data payload, it contains a callback address and unique identifier. This, in essence, is what the WS-Addressing standard is about.
Now the definition of a one-way invocation type in a WSDL is:
    <portType name="aaa">
        <operation name="bbb">
            <input message="tns:ccc"/>
        </operation>
    </portType>Compare that to a request-response invocation type:
    <portType name="aaa">
        <operation name="bbb">
            <input message="tns:ccc"/>
            <output message="tns:ddd"/>
        </operation>
    </portType>Let's look again out our example WSDL:
    <portType name="SelectService">
        <operation name="processRequestQuote">
            <input message="tns:RequestQuote_processRequestQuote"/>
        </operation>
    </portType>
    <portType name="SelectServiceCallback">
        <operation name="processRequestQuoteResponse">
            <input message="tns:RequestQuote_processRequestQuoteResponse"/>
        </operation>
    </portType>Here comes the good bit... Both portTypes have an <input> operation. But that's because they are both one-way invocation types, and there is no choice but use the <input> element -- that's the standard. You can't imply put <output> because there's no such thing in the standard. However we know that one of those is to actually return the result. That is, it's the output, even though it's labelled <input>.
Hopefully that's given you enough information now. Re-read my first post, above, and it should make more sense.
Incidentally, this is why you rarely see try asynchronous web services, because the caller has to also be a listener. And if you want to call a service, who wants to also have to write code to listen, to handle responses coming out of order, etc. This is one of the advantages of using an orchestration engine like Oracle BPEL Process Manager. The framework takes care of the hard work, and you can simply call an aysnchronous service and not have to worry about how to get the response back -- the engine does it for you.
Regards,
Robin.

Asynchr Calls: Partner Role and My Role must be filled in Partner Link dial

Asynchr Calls: Partner Role and My Role must be filled in Partner Link dialog ?
Is the statement true?
Can I leave for synchronous calls one of the role spec vacant?
AFAIK only in asynchronous calls both drop down values must be specified.
Peter

Thanks for fast reply,
Yes, I defined Partner Role and My Role,and also defined a port type for calling back the request.
I just don't know what's the endpoint address which my asynch web service should use to response to my BPEL process.
Is there anyway to find out what's the endpoint address which the BPEL processes receive the callback response from asynchronous services?
Can I get the address which BPEL engine uses to get the responsse from a called web service?
I guess it should be an obvious work but I'm really disappointed with that.

Funktion Roles and Value Roles

Hello,
i read in a SAP Press book something about funktion roles and value roles.
Can someone explain me how this work.
kind regards,
Bernhard

I would suggest taking preventative legal action against anyone who even mentions "functional and value" roles - particularly if they give the impression that transaction codes, activities and org-levels can be built in seperate roles - because when the concept goes downhill (which it will!) then they will unlikely be around to clean up the mess nor take responsibility for it.
Rather steer well clear of this type of concept.
Cheers,
Julius

Business Role and PFCG Role

Hi all,
I am new to CRM 7.0 Can someone explain What is a Business Role in CRM 7.0 and what is the relationship between Business role and PFCG role. What is the transaction Code to create a Business role.
And also I heard that there is no PCUI in CRM 7.0. Is it true and if so what is used in place of the PCUI
Thanks.
Neha.

Neha,
Next time please do a search in this forum on business roles, and you would find many topics discussing this information more completely. I'm locking this thread due to it fact that this question has been asked many times before by many different people.
These threads explain the topic in more detail:
Re: Reg: Business Role
Assignment pfcg-role to user and assignment pfcg-role to business role
Thank you,
Stephen

Role and Reverse Role

Can the Role and Reverse Role fields be removed from the account competitor view? Thanks.

You can hide the Role fields in the Account Related Information Layout, which a user will see from the Account view. However, the fields are required, so when creating or editing a related account, users will have to see the fields. You can pre-default the values to make it easier for users, though.
Good Luck,
Thom

DB table for Derived Roles and Parent Roles

Hi Expart,
In which DB table the Derived Roles and Parent Roles are store .that is i need to find out the derived role and parent Role .i have completed the Complex and single role by table AGR_AGRS
But i have to find out the table for Derived Role
Plz help me to get those table
Thanks in advance
Tarak

It's the same table as for the master role: AGR_DEFINE (field PARENT_AGR is filled for derived roles).
~As from Forum

Security report with native roles and the roles they have access to.

We need a security report that shows the Native/Custom Roles and the roles that they have access to.
So, an example would be the role US_Acct, and the report would show what roles that has access to (Post Journals, Consolidate, etc).Can this be done?

Export the Provision report from Shared Services.
Upload report to Excel or Access.
Build Tables to show what tasks each Role has access to.
Build a report that links the provision report and the xref tables.
You should also do this with Security Classes.

Maintaining the authorizations for parent role and derived role

Hi Experts,
Kindly advice me the Pro and cons of the parent role and derived role.. below is the scenario
Currently we have created the 700 role in our regionally organization and we want to dervie the roles for each country
1 ) we want to do the Auth field (activity level) settings in parent role and Org levels in the derived role .
2) But one my collegue says do the default Auth filed ( activity values) common to every country in the parent role and diff activity one in the derived role .
please advice me wat will be the best scenario for mantaining the authorizations filed values like (activity level one)

I will try to answer both your queries here:
"my collegue says they are some NON ORG values different from each country ..suggest us to maintain all the default values in Parent role and auth with diff values needs to be maintained in derived role (child role).. "
The only set of values which should/can be different in a child role (when compared with its parent) will be the org level values. So if this filed is NON_ORG you will not be able to maintain it directly inside the child roles.....this is the basic principle of derived role conceptu2026 that the only item you will directly maintain in a child role are the org levels(which will come as u2018organisational levelsu2019 in the upper tab in the auth data of a role).
All NON_ORG fields inside a child role is acquired from the parent role. You should never change the values of any such fields (non-org fields) in the child role. these changes will get lost the next time you run the parent child inheritance from u201Cgenerate derived roleu201D function in your parent role.
Coming to the second question on how to run the program, you just need to enter the technical name of the field you want to convert (tech names like BUKRS, WERKS etc u2026 figure out the name of the concerned field you have in hand)u2026.executeu2026 you will that the field will now onwards appear as an org level value in all roles in the system and not just as a field inside the auth objectsu2026.I would suggest you take one field and try running it in ur dev or sandbox..see how the field changes in your roles.... the change can always be reverted by using PFCG_ORGFIELD_delete. ... you will understand it better....
Soumya

BeX-Analyzer : no visible role and working-sheets

Hello experts,
after creating new working-maps under a role, i can't find it anywhere for the next day.
For example: After launch BeX-Analyzer 7.0 -- then open working-map -- then connect over BW/BI Developer-System -- then
roles. I can't find my saved user-working-maps or local objects who are saved one day before. They were under the "PFCG-created"
user-role : "example :Z_Role_External_".
When i open the map : QUERIES, i have the include for user-role and can manage and call the designed sheets
in my workaround.
In this case one more information that happens before this items. In the last week i had patched the BW-BI 7.0 with then following patches.
SAPKA70020
SAPKB70021
SAPKA70021
KIPYJ7K und KIPYJ7L
SAPKW70022
SAPKW70023
After this action i heard from this problem in BW-BI "No User-Role and no Test-working-maps" under my User-Profile.
Have anybody an idea to solve this problem ?
Thanks
Michael

Found this thread with an allegeded fix from SAP
Unable to see Roles in Bex Analyzer - even with SAP_ALL!

BeX-Analyzer : no visible role and working-sheets created bei querie

Hello experts,
after creating new working-maps under a role, i can't find it anywhere for the next day.
For example: After launch BeX-Analyzer 7.0 -- then open working-map -- then connect over BW/BI Developer-System -- then
roles. I can't find my saved user-working-maps or local objects who are saved one day before. They were under the "PFCG-created"
user-role : "example :Z_Role_External_".
When i open the map : QUERIES, i have the include for user-role and can manage and call the designed sheets
in my workaround.
In this case one more information that happens before this items. In the last week i had patched the BW-BI 7.0 with then following patches.
SAPKA70020
SAPKB70021
SAPKA70021
KIPYJ7K und KIPYJ7L
SAPKW70022
SAPKW70023
After this action i heard from this problem in BW-BI "No User-Role and no Test-working-maps" under my User-Profile.
Have anybody an idea to solve this problem ?
Thanks
Michael

Hi,
this forum is about Web Intelligence. I would suggest you post your questions into the BusinessExplorer (BEx) Forums.
ingo

Scaling of VM Roles and Worker Roles, why divide by total?

Similar Messages

Maybe you are looking for