Wednesday, June 14, 2017

Production instances facing the Public

Here production instances facing the public.
Admin works:
Setup of instances and environments:
—————————————————————————————————
extra points:IP address to domain  mapping (networking guy).
—————————————————————————————————-
The figure of 123 is called 123 architecture
in that :
one author
2 publish(another one will be in standby).
2 dispatchers.
——————————————————————————————————
what is meant by replication agent?
1)whenever you want to configure the author to publish.(i need to configure the endpoints in my replication agents that is done by the author).
2)what is dispatcher?
Dispatcher is just a module from Adobe,when you install any webserver,you will get generated with the folder called module.you need to go and place the module under the webserver.the moment we do this your webserver is taking to dispatcher.
(all the end user request, will be intercepted by your dispatcher one is caching and another one is load balancing)and also you dont want to expose your server directly that is why we are doing this.
———————————————————————————————————————————————————
For example your dealing with website, the request  hits the CDN(content delivery network):my enterprise is distributed across different geolocations for example A,B,C  are different locations ,if you get more requests from C ,then the requests will not be fulfilled from the location where AEM servers are present ,it will fulfilled(or redirected to that C servers) at the C location,here are the some of the providers for CDN’s:Akamai,china cache and amazon has there own CDN,the moment you subscribe this CDN providers, any request that comes to that host will be redirected to that servers to fulfill the request with the help of CDN providers.which means whatever there in my AEM ,that means ideal whatever the request comes will go to the akamai ,the akamai will cache the things from the Dispatcher,so that way whenever there is a request from that particular location, in that location akamai will already have the cache from dispatcher,willserver the requests, so in this way you no need to communicate with the AEM dispatcher in order to serve the request
Benefits of using this akamai(that is CDN (content delivery network):
     we are able to avoid the latency
     we are able to avoid the network dependencies. and then loadbalancer(LB),
     you need not to wait for the response from that particular server.(CDN will pick the response from the nearest geolocation and serves the response).
CDN is a separate service, it is mostly cloud service.they do caching also.
Diff between CDN and Dispatcher:
With CDN latent issues are avoided, by avoiding that, we can improve the performance.

example:If the request is for the first time,If CDN is not cached anything ,the content will go to the dispatcher and then control will come to the load balancer.
They are two kinds of Load balancer:Physical load balancer and software load balancer .
usually we use the software load balancer, which means there will be the random alogarithm,some of the famous load balancer is the BigIP.
The dispatcher is getting content from the publisher.
————————————————————————————————————————————————————coming to flushing agent:
Dispatcher can do the caching also as we know we are using different mechanisms to flush the things, as we know akamai is the configurable tool,akamai acts like a cache layer to the dispatcher.we have to ensure that akamai cache and dispatcher cache are in sink.Mostly wildly used is akamai.(typically it may be a webserver).we dont know exact thing what they are using, we need to refresh the content, when akamai is required, there will be akamai controllers to take care of this.
Note:we just need to provide the content and we need to provide the path of the dispatcher path, where the content is stored.
we need to go to built with: and in search bar provide the site name and that provides the usage of the CDN.
we can put google.com in search bar and we can know which CDN there are using.
——————————————————————————————————————————————————————
1)we need to setup the author, publish and dispatcher instances
we need to setup the load balancer.
we need to explain the setup related to dispatcher to the CDN providers about the content.
2)Configuring the replication agents.
a)Replication agents
b)reverse replication agents
c)Dispatcher flushing agents
3)Diagnosing and maintaining instances
e.g.: assume server has gone down and (something called health check will take care of that)
you need to go on run the health check and find out anomous either in OSGI bundles or some of your queries are taking a lot of time let us take it is taking 5 to 10 mins to execute (these things will hit on your performance) this is nothing but the performance tuning (that means your also responsible for performance tuning).
4)Responsible for the security.
(Initially when your setting up, there are few points which they mention as security checklist:
(this is done by admin.)
5)We are completely responsible for maintaining of logging.
——————————————————————————————————————————————————
1)Rotation mechanism
Timestamps
For entire environment configuration, that is different environments have different configurations (logging configuration)the owner of entire thing is the admin.

2)The most important point of maintainence:
Different maintainence task:
1)Backup
2)Tar optimization
3)logs
4)compassion

3)User administration:
(the owner of production will become the administration)
Authorizations on different paths
New userid additions
4)Upgrade activities
whenever adobe releases the new patch it is called service packs (SP)
(like 5.6 to 6.1)
or
some times may be the hot fix (that fix is related to the feature.
(the responsibility of the admin to ensure the releases).
—————————————————————————————————————————————————————
Some tasks:
Who sets here the LDAP??
whenever u want to add the new user it should relay on the LDAP,but here the AEM admin have that feature of LDAP framework.
(It is a job of an admin to configure the LDAP)
—————————————————————————————————————————————————————
What is meant by SSO??
it is single sign on
It is single userid and password for all login.
this is possible when you have an LDAP.
——————————————————————————————————————————————————
Our enterprise can also have called SAML.
we are associating this with the help of SSO.
(Security  accessing markup language)
———————————————————————————————————————————————————
Other things:
Infrastructure:
You also need to profile your JVM:
1)you have thread dumb
2)you have heap dumb
(by using this heap dumb you can exactly know the usage of memory)
———————————————————————————————————————————————————
Whenever you install anything:
1)You have to know the different ways of installing
a)Jar+license (we need to create a folder called the folder as author, adobe has a different name convention that is cq-instance type-port<450x)
Author we have to name as even numbers
Publish have to name as odd numbers
cq-instance type - port <450x>
Publish have to name as like this.
a)double click on the jar file
b)we can from the command line
java -jar cq-author-port number.
if we are facing heap exception problems we have to use this command:
java -Xms :256mb -jar cq-author-portnumber 
java -Xmx:1024 -jar cq-author-portnumber
In addition the entire JVM  you would have a reflective data of entire JVM most of them are outofmemory errors because of JVM      -Xx permsize 
Generally we have a memory area called permanent generation  which needs to be more than 1024 mb, when ever you get out of memory problem blindly we can go and increase the perm size.
what is meant by reflective data??
it consists of all your class, functions informations so that is why is called reflective data.
Usually in production we use the 10gb
when your starting the server, you can enter the complete information of
java  xxxxxx)size of the perm size) and jar instance ,port number.
c)Batch file (CRX-quick start  ————>start t
he batch file that is start.bt)
Otherwise the tuning setting can also be given by go to the CRX-quickstart > bin >  start.bat (give all the port configurations and another configurations the settings,configure the run mode ,now the server is started.

whenever you feel that they is a change in product view definitely it is a change related to the osgi bundle configuration.

we can change the default one from the OSGI configuration that is nothing but the root mapping .
Note:Whenever there is a behavioral change, there is no point of code, you just need to go to the OSGI config files and check the identity.
—————————————————————————————————————————————————
what is meant by proxy??
whenever i have server called proxy,
whenever i am hitting the url,instead of directly hitting the server ,the request will go to the proxy.
[what is demilitized zone?(DMZ)
your entire production will be demilitized zone(DMZ)
All production servers are in demilitarized zone.
what is meant by firewall?
which will block all the unwanted requests.
when i am configuring servers obviously it will be behind the firewall.
Note:If you have a firewall, why do you require DMZ??]
whenever you have the proxy you need to bypass all the requests,
for this there is a configuration, to know about the enterprise level proxy settings we need to contact the network team.
—————————————————————————————————————————————————

Every enterprise have it’s own SMTP for email notification services.

No comments:

Post a Comment