Chandra A Sekhar

Greenhorn
+ Follow
since Dec 11, 2000
Merit badge: grant badges
For More
Cows and Likes
Cows
Total received
In last 30 days
0
Forums and Threads

Recent posts by Chandra A Sekhar

We are trying to implement a system in which we have Course and Student Objects, each student randomly registers in to some course however each course has a constraint that it can only accommodate 16 students,if it exceeds we will create a new course instance and start populating the new Course object.
We are worried about using any Singleton Pattern for this kind of system as we expect our system has to scale down for a large traffic which means we have to use clustering for appservers. We know for sure that in a clustered environment we can not use this singleton pattern as each JVM will end up with their own version of it.
One solution we came up with is using a JMS queue and making our CourseManager(Singleton) to listen to this JMS queue so that any update on one CourseManager instance is propageted to other Clustered instances via JMS.
Is there any other simpler way we can solve this problem of keeping two objects in a clustered environment in sync with each other ? or for that matter is there a way we can implement singleton across clustered servers ?
Any input is well appreciated,
Thanks in advance
Chandra
Hi,
I have a general design question ,we have some webservices and their clients which post some data to webservices. This whole process is triggered when user clicks somebuttons during screen flow. But the whole thing is asynchronous ie we just need to trigger the clients and after that the flow doesnt bother about message delivered to webservices immediately or not.(but eventually it must be delivered at some point of time)
If the services is up and running posting this client request to webservice is stright forward, however if the webservice is not available some reason then the ideal mechanism is kind of generate the soap request message and que them and post them to the webservices when it is available , and for obvious reasons we need to make sure that we keep trying theoritically forever till the service is available again.
We are just trying to come with some common frame work which we can use across all our applications which talk to webservices. One idea is to use a JMS server and as soon as we generate soap message we que them in JMS server and a local client which is polling on to this JMS server que tries to post messages (if available)to the webservices if it is successful it will clear the message from JMS server otherwise it will not clear and this process repeats for ever.
Is there any other simple model/framework/mechanism is alreay in place to asynchronously sending soap requests ? and at the same time ensuring that eventually the message gets delivered.?
I heard some message servers even allow us to configure to post to a http client so that we can directly configure the server to post to the end webservices leaving the posting business to the jms server it self , but at the same time ability to post to http clients it is not mandated by JMS specification ? so we may loose the ability to port our code to other jms servers ?
Any ideas comments ?
Thanks in advance.
Chandra
22 years ago
Hi Syed,
Thanks for explanation and reply, but I dont have that book with me to refer, however I could get that code Iam talking about .

*************************************************

*************************************************
Ok this is part of the DTD I designed , and we have xml instances which are validated against this DTD using XERCES. Here Topic can contain UNIT,any of the peripheralitems or QUIz in any order . Just now I again tried by giving random order of elements Xerces validates the document and as does Turbo Extensibility.
Iam not sure about the exact rules , but I think by this way we can define that elements can appear in any order.
Any DTD guru's around here who can thro some light on this topic ?
Thanks in advance
Chandra
Hi Guys,
First of all thanks for the link !, One of the question here says we cannt define that elements can appear in any order using DTD semantics ( the question says we can do that only with Scema's ) . Is this true I vaguely remember I did that using DTD in one of my early projects but now I forgot ! anyone has a better Idea about it ? Can we define that elements can appear in any order using DTD ?
Thanks in advance for any input
Chandra
Hey John,
Got it ! Thanks for the explanation and reference.
Chandra
Quiz 2 in perfect xml :
The following XML document is legal:
<!--Document last updated: January 15-->
<?xml version="1.0" ?>
<SalesReport year="2001" month="January">
<NetIncome>25,000</NetIncome>
<Profit>2,000</Profit>
</SalesReport>
True False
The answer it says is false , I couldn't find any thing wrong with this structure , is there anything Iam missing?
Thanks in advance
Hi,

We have a strange problem, we have 9 different threads accessing one table (ofcourse thro a synchrnized DBClass method) , each time each thread needs around 2-3 mins and thread will come back in 30 mins. In other words this is more of a continuous process.
Once in a week we need to clean up the complete table during that process our clean up procedure will archive all previous weeks data on to archive data base.
My question here is the archivel process takes some time (about 30 mins) during that process my 9 threads will try to write on to that table and they will keep waiting for clean up procedure to leave the lock , so here can anybody tell me how much time will that connection be valid and how much time it waits or keeps polling for a write access on to that table ?
Some times all 9 threads die at this very particular time so we are suspecting that some thing is going wrong during thar process.!!
Any input will be greatly appreciated

Thanks in advance,
Chandu
Thanks Matt, When I registered for the first time this rule was not there . Anyway now am done.
24 years ago