Search Results

Search found 26810 results on 1073 pages for 'fixed point'.

Page 532/1073 | < Previous Page | 528 529 530 531 532 533 534 535 536 537 538 539  | Next Page >

  • How to use shared_ptr for COM interface pointers

    - by Seefer
    I've been reading about various usage advice relating to the new c++ standard smart pointers unique_ptr, shared_ptr and weak_ptr and generally 'grok' what they are about when I'm writing my own code that declares and consumes them. However, all the discussions I've read seem restricted to this simple usage situation where the programmer is using smart in his/her own code, with no real discussion on techniques when having to work with libraries that expect raw pointers or other types of 'smart pointers' such as COM interface pointers. Specifically I'm learning my way through C++ by attempting to get a standard Win32 real-time game loop up and running that uses Direct2D & DirectWrite to render text to the display showing frames per second. My first task with Direct2D is in creating a Direct2D Factory object with the following code from the Direct2D examples on MSDN: ID2D1Factory* pD2DFactory = nullptr; HRESULT hr = D2D1CreateFactory(D2D1_FACTORY_TYPE_SINGLE_THREADED, &pD2DFactory); pD2DFactory is obviously an 'out' parameter and it's here where I become uncertain how to make use of smart pointers in this context, if indeed it's possible. My inexperienced C++ mind tells me I have two problems: With pD2DFactory being a COM interface pointer type, how would smart_ptr work with the Add() / Release() member functions for a COM object instance? Are smart pointers able to be passed to functions in situations where the function is using an 'out' pointer parameter technique? I did experiment with the alternative of using _com_ptr_t in the comip.h header file to help with pointer lifetime management and declared the pD2DFactory pointer with the following code: _com_ptr_t<_com_IIID<pD2DFactory, &__uuidof(pD2DFactory)>> pD2DFactory = nullptr; and it appears to work so far but, as you can see, the syntax is cumbersome :) So, I was wondering if any C++ gurus here could confirm whether smart pointers are able to help in cases like this and provide examples of usage, or point me to more in-depth discussions of smart pointer usage when needing to work with other code libraries that know nothing of them. Or is it simply a case of my trying to use the wrong tool for the job? :)

    Read the article

  • Weird routing problems with VPN

    - by Borek
    In our VPN setup I have to add a route to my routing table like this: route add 1.2.3.0 mask 255.255.255.0 172.16.1.1 -p Our internal addresses 1.2.3.x then use 172.16.1.1 as their gateway and both my local internet and work VPN can work at the same time. However, when I disconnect from VPN and reconnect again, I can't ping our servers even though the connection status is "Connected". When I do route print my previously added route is listed but it doesn't seem to work. So I try to execute that 'route add' command again and as expected, it tells me that The route addition failed: The object already exists. But - and that's the point - when I now try to ping our servers again, everything works! So every time, I have to execute this route add command that will fail but fix the issue at the same time. Any ideas what I might be doing wrong? My PC is Windows 7 x64, I am Administrator, UAC is enabled and the command prompt is run with elevated privileges.

    Read the article

  • video streaming infrastructure advice

    - by Alchemical
    We would like to set-up a live video-chat web site and are looking for basic recomendations for software and hardware set-up. Most streams will be broadcast live from a single person with a web cam, etc., and viewed by typically 1-10 people, although there could be up to 100+ viewers on the high side. Audio and video do not have to be super-high quality, but do need to be "good enough". The main point is to convey the basic info in the video (and audio). If occasionally the frame-rate drops low and then goes back to normal fairly soon, we could live with that. Budget is an issue, so we are in general looking for a lower cost solution that will give us most of what we need in temers of performance and quality. We are looking at Peer1 for co-lo. The rest of our web site will be .Net / Windows platform. We are open to looking at any platform for the best streaming solution, although our technical expertise is currently more on the Windows side.

    Read the article

  • A question about the cobbler-ubuntu-import bash script

    - by user183394
    I have been testing the latest cobbler for PXE booting Ubuntu 12.04.1-server-x86_64 and 12.10-server-x86_64 using a Scentific Linux 6.3 host to run the cobbler server. With the former, I got everything going. But with the later, I haven't been successful. As an attempt to figure things out, I downloaded Ubuntu's cobbler 2.2.2 source package. Examining the content, I soon noticed that Ubuntu's cobbler 2.2.2 came with a cobbler-ubuntu-import bash script. I reviewed the code and spotted something interesting: line 9 of the script states: 9 AUTO_KOPTS='log_host=@@server@@ log_port=514 priority=critical locale=en_US netcfg/choose_interface=auto' But after extensive googling, reading both Debian and Ubuntu's documentation about the debian-installer, I don't see these two kopts log_host and log_port documented anywhere. Putting it in the profile of my current test setup, even my cobbler server host does run rsyslogd, I don't see anything logged either. No, I don't have iptables and selinux on on the cobbler server host. Can anyone point to me where I can read more about these two options? Having the ability to log an installation to a remote central logging host would be really cool.

    Read the article

  • Can I get "disk utilization" from a NetApp filer via SNMP?

    - by Andrew
    On a NetApp filer's command line I'm running "sysstat -u" to show disk utilization, (actually the utilization of the single busiest disk). By disk utilization, I mean "percent of time the disk is busy", not "how much space on the disk is being used to store data/metadata". Is there a way to get disk utilization info through SNMP? The netapp.mib file doesn't appear to expose this. It does have CPU utilization, disk usage & capacity information, etc, but not disk utilization. The MIB-II (rfc1213) seems to be the only other information exposed by the filer through SNMP. I hope I am missing something. The "CP (consistency point) time" metric is exposed through the NETAPP-MIB in SNMP, but this seems to only partially correlate with disk utilization under write load, and not really at all under read load.

    Read the article

  • Bad at math, feeling limited

    - by Peter Stain
    Currently I'm a java developer, making websites. I'm really bad at math, in high school I got suspened because of it once. I didn't program then and had no interest in math. I started programming after high school and started feeling that my poor math skills are limiting me. I feel like the programming's not that hard for me. Though web development in general is not that hard, i guess. I've been doing Spring and Hibernate a lot. What i'm trying to ask is : if I understand and can manage these technologies and programming overall, would it mean that I have some higher than average prerequisite for math and details? Would there be any point or would it be easy for me to take some courses in high school math and get a BSc in math maybe? This web development is really starting to feel like not my cup of tea anymore, i would like to do something more interesting. I'm 25 now and feel like stuck. Any help appreciated.

    Read the article

  • Installing Apache2 and PHP5 on a XServe G5 running 10.4

    - by Chris
    Hello- I have an Apple XServe G5, running OS X Server 10.4. I want to update my Apache installation from 1.3 to 2 and PHP to 5. I also want to install PHP GD support. I have scoured the internet for a guide on how to do this, but to no avail. I also tried to use Entropy to install PHP 5 several times, and it always manages to royally mess up my system. Obviously I can't install Leopard or Snow Leopard because it is a PowerPC processor. Can anyone give me any tips on how to get this software updated or point me to a guide? Thanks, Chris

    Read the article

  • Why does my system use so much cache?

    - by Dave M G
    Previously, on my desktop computer running Ubuntu 14.04, I had 4GB RAM, which I thought should be plenty. However, after being on for a while, my computer would seem to get slow. I have a system resource monitor app in my Gnome panel, which I assume represents the available RAM (?). It shows a dark green area as being "Memory", and a light green area as "Cache". The "Cache" would slowly grow until it filled the whole graph, and then programs would get slow to load, or it would take a while to switch programs. I could alleviate the problem somewhat with this command, but eventually the computer cache fills up again, so it's only a bandaid: sudo sh -c "sync; echo 3 > /proc/sys/vm/drop_caches" So, I figured I'd get more RAM, so I replaced one 2GB stick with an 8GB stick, and now I have 10 GB ram. And my "cache" still slowly maxes out and my computer slows as a result. Also, sometimes the computer starts out with "cache" maxed when I first boot and log in. Not always though, I don't know if there's a pattern that determines why it happens. Why is Ubuntu using up so much cache? Is 10GB not enough for Ubuntu? Here's what my system monitor looks like in my Gnome panel. The middle square shows RAM usage. The light green area is the "cache": This is my memory and swap history, which doesn't seem to include any information about "cache". I realize at this point I'm not totally clear on the difference between "cache" and "swap":

    Read the article

  • I need something substantial to do [on hold]

    - by Christian Delapena
    I'm a 19 year old computer geek who was recently exposed to Linux. I know quite a bit of it now and would like to do something substantial with it. I've visited websites like openhatch where you can get started on opensource projects but I'm more interested in something Linux-specific like hosting a website or tracking some important operation. Maybe running a script that will give me data on something important. I don't know. I'm essentially bored and I want to put my knowledge and love of Linux to good use. Someone please point me in the right direction.

    Read the article

  • SQL SERVER – Technical Reference Guides for Designing Mission-Critical Solutions – A Must Read

    - by pinaldave
    Yesterday I was reading architecture reference material helping my friend who was looking for material in this respect. While working together we were searching twitter, facebook and search engines to find relevant material.While searching online we end up on very interactive reference point. Once I send the same to him, he replied he may not need anything more after referencing this material. The best part of this article was it gives access to various aspect of the technology of the image map. Here is the abstract of the original article from the site: The Technical Reference Guides for Designing Mission-Critical Solutions provide planning and architecture guidance for various mission-critical workloads deployed by users. These guides reflect the knowledge gained by Microsoft while working with customers on mission-critical deployments. Each guide provides not only the key technical concepts and information helpful for design, but also “lessons learned,” best practices, and references to customer case studies. Once you click on any of the desired topic, you will see further detailed image map of the selected topic. Personally once I ended up on this site, I was there for more than 2 hours clicking through various links. Click on image to see larger image Read more here: Technical Reference Guides for Designing Mission-Critical Solutions Reference: Pinal Dave (http://blog.sqlauthority.com) Filed under: PostADay, SQL, SQL Authority, SQL Query, SQL Server, SQL Tips and Tricks, SQL White Papers, T SQL, Technology

    Read the article

  • Cocos2d sprite's parent not reflecting true scale value

    - by Paul Renton
    I am encountering issues with determining a CCSprite's parent node's scale value. In my game I have a class that extends CCLayer and scales itself based on game triggers. Certain child sprites of this CCLayer have mathematical calculations that become inaccurate once I scale the parent CCLayer. For instance, I have a tank sprite that needs to determine its firing point within the parent node. Whenever I scale the layer and ask the layer for its scale values, they are accurate. However, when I poll the sprites contained within the layer for their parent's scale values, they always appear as one. // From within the sprite CCLOG(@"ChildSprite-> Parent's scale values are scaleX: %f, scaleY: %f", self.parent.scaleX, self.parent.scaleY); // Outputs 1.0,1.0 // From within the layer CCLOG(@"Layer-> ScaleX : %f, ScaleY: %f , SCALE: %f", self.scaleX, self.scaleY, self.scale); // Output is 0.80,0.80 Could anyone explain to me why this is the case? I don't understand why these values are different. Maybe I don't understand the inner design of Cocos2d fully. Any help is appreciated.

    Read the article

  • Export GFI MailArchiver e-mails for import into Exchange 2010 SP1 Personal Archiving

    - by pk.
    We have an existing installation of GFI MailArchiver 5 with several databases of archives (perhaps 100-150GB). The goal is to export each user's archived e-mail and then import it into Exchange 2010 SP1 Personal Archives. GFI has a tool to do this, but it's very rudimentary and has severe, frankly unworkable, limitations. It only allows me to query based on the e-mail headers. Due to the fact that we have multiple aliases that may show in multiple headers (To:, Cc:), not to mention the fact that this won't cover a user's membership in a distribution group at a given point in time, this tool will not suffice. Another option is for me extract the e-mails from the GFI databases without using the tool, but this would require me to write my own tool to reconstruct them and I really would rather not go down that path. I feel very stuck on this issue. Has anyone here done a similar migration? How can this best be handled?

    Read the article

  • Is rotating the lead developer a good or bad idea?

    - by Renesis
    I work on a team that has been flat organizationally since it's creation several months ago. My manager is non-technical and this means that our whole team is responsible for decision-making. My manager is beginning to realize that there are several benefits to having a lead developer, both for his sake (a single point of contact and single responsible party for tasks) and ours (dispute resolution, organized technical guidance, etc.). Because the team has been flat, one concern is that picking one lead developer may discourage the others. A non-developer suggested to my manager that rotating the lead developer is a possible way to avoid this issue. One developer would be lead one month, another the next, and so on. Is this a good idea? Why or why not? Keep in mind that this means all developers — All developers are good, but not necessarily equally suited to leadership. And if it is not, suppose I am likely the best candidate for lead developer — how do I recommend that we avoid this approach without looking like it's merely for selfish reasons? (In other words, the team is small enough that anyone recommending a single leader is likely to appear to be recommending themselves — especially those who have been part of the team longer.)

    Read the article

  • Economical DNS hosting separate from local registrar for country specific TLDs - email or web hosting not required

    - by Eric Nguyen
    Our company owns many country specific top level domains (TLDs; .sg, .my). We will purchase more for other countries in all South East Asia. These domains are associated with our websites hosted on Amazon EC2. The DNS records are currently hosted on a dedicated server that will shut down tomorrow. (The name servers are set to the ones of a web hosting company) Therefore, I will need to host the DNS records somewhere else. Hosting the DNS records with the local registrar costs SGD18 a year per domain in addition to the domain price (which is already very expensive but we have no choice). It would be convenience to host DNS recors for all the country specific TLDs we have using a single service, separate from the local registrars from which we bought the domains. A few searches prompted examples like Amazon Route 53 and dnsmadeeasy.com and the likes. However, since I'm only concern about the country specific TLDs, not .com 1) Is it really economical to host DNS records of all domains in 1 single place as described above? (Have the relevant countries and/or the local registrars done something to keep their monopoly and always charges ridiculous prices for their country specific TLDs?) 2) I would imagine I will need to tell the local registrars to update the name servers to those of the DNS hosting service provider e.g. dnsmadeeasy.com here. Am I correct about how it works here? 3) Will I be able point the TLDs themselves to IP addresses I desire (the EC2 instances where my websites are) or will I only able to do so with the subdomains? 4) Are there any drawbacks that I should know here? Background about our needs: We need the websites associated with the country TLDs to be up and running all the times Also, we'll need to be able to add/edit A and CNAME records We use Google Apps for Business for internal email so I will need to be able to add/edit MX records and TXT records

    Read the article

  • Is it wise to ask about design decisions made on a product during an interview?

    - by Desolate Planet
    I've been thinking about interview questions lately and I've been reflecting on bad interview experiences I've had in the past. One of particular note is where I had asked the interviewer why the team chose to use Spring over EJB3 in their product. The interviewer pretty much tore my face off, yelling "Because Spring is not the be all and end all of Java software development, do you want this job or not?". In response to this, I told him that this probably wasn't the job for me and I walked out the interview. He told me at the start of the interview that they had high stuff turnover, the product had gone from Modula 3 to Perl to Java then after asking him a technical question, he went in flames. It seemed obvious to me that he was toxic to the company with that kind of attitude. Question: Is it a good idea to probe on architectural choices taken in an interview? If not, why? From my own point of view, an interview is a two-way process. If the interviewers are testing me on my technical skills, I've got every right to ask them the same questions to 1) Figure out what their mindset and attitudes towards developing software solutions are and 2) To figure out if there are in line with how I would approach problems of that kind. It's very possible that the interviewer who got angry was a bad interviewer and forgot that an interview is a two-way process. If I was asked this, I would have simply said something along the lines of wanting to leverage the container more, but I certainly wouldn't have tried to put him in a state of meek capitulation. The interviewer in question was the lead developer in the team.

    Read the article

  • Is it possible to redirect/bounce TCP traffic to external destination, base on rules?

    - by xfx
    I'm not even sure if this is possible... Also, please forgive my ignorance on the subject. What I'm looking for is for "something" that would allow me to redirect all TCP traffic arriving to host A to host B, but based on some rules. Say host A (the intermediary) receives a request (say a simple HTTP request) from a host with domain X. In that case, it lets it pass through and it's handled by host A itself. Now, let's suppose that host A receives another HTTP request from a host with domain Y, but this time, due to some customizable rules, host A redirects all the traffic to host B, and host B is able to handle it as if came directly from domain Y. And, at this point, both host B and the host with domain Y are able to freely communicate (of course, thought host A). NOTE: All these hosts are on the Internet, not inside a LAN. Please, let me know if the explanation is not clear enough.

    Read the article

  • Should we write detailed architecture design or just an outline when designing a program?

    - by EpsilonVector
    When I'm doing design for a task, I keep fighting this nagging feeling that aside from being a general outline it's going to be more or less ignored in the end. I'll give you an example: I was writing a frontend for a device that has read/write operations. It made perfect sense in the class diagram to give it a read and a write function. Yet when it came down to actually writing them I realized they were literally the same function with just one line of code changed (read vs write function call), so to avoid code duplication I ended up implementing a do_io function with a parameter that distinguishes between operations. Goodbye original design. This is not a terribly disruptive change, but it happens often and can happen in more critical parts of the program as well, so I can't help but wondering if there's a point to design more detail than a general outline, at least when it comes to the program's architecture (obviously when you are specifying an API you have to spell everything out). This might be just the result of my inexperience in doing design, but on the other hand we have agile methodologies which sort of say "we give up on planning far ahead, everything is going to change in a few days anyway", which is often how I feel. So, how exactly should I "use" design?

    Read the article

  • How can I join two simple home networks together using an ethernet cable?

    - by Ilia Jerebtsov
    I want to join two different home networks together like so: PC A1 PC A2 PC B1 PC B2 \ / \ / Gateway A <----- ethr. cable -----> Gateway B | | ADSL modem A ADSL modem B Both networks are of the basic residential type with identical configuration, with all PCs running Vista/7. The point is to temporarily join two apartments in a building for gaming and file sharing, and the holy grail would be: PCs on network A can access PCs on network B and vice-versa (file shares and gaming). Each network uses its own internet connection. Data between networks shouldn't take a trip through the internet (broadband upload speeds are severely capped) A network's internet access should continue working if the joining cable is disconnected with minimal configuration changes. How closely can this be achieved?

    Read the article

  • How can I rank teams based off of head to head wins/losses

    - by TMP
    I'm trying to write an algorithm (specifically in Ruby) that will rank teams based on their record against each other. If a team A and team B have won the same amount of games against each other, then it goes down to point differentials. Here's an example: A beat B two times B beats C one time A beats D three times C bests D two times D beats C one time B beats A one time Which sort of reduces to A[B] = 2 B[C] = 1 A[D] = 3 C[D] = 2 D[C] = 1 B[A] = 1 Which sort of reduces to A[B] = 1 B[C] = 1 A[D] = 3 C[D] = 1 D[C] = -1 B[A] = -1 Which is about how far I've got I think the results of this specific algorithm would be: A, B, C, D But I'm stuck on how to transition from my nested hash-like structure to the results. My psuedo-code is as follows (I can post my ruby code too if someone wants): For each game(g): hash[g.winner][g.loser] += 1 That leaves hash as the first reduction above hash2 = clone of hash For each key(winner), value(losers hash) in hash: For each key(loser), value(losses against winner): hash2[loser][winner] -= losses Which leaves hash2 as the second reduction Feel free to as me question or edit this to be more clear, I'm not sure of how to put it in a very eloquent way. Thanks!

    Read the article

  • Where could Distributed Version Control Systems currently be in Gartner's hype cycle?

    - by dukeofgaming
    Edit: Given the recent downvoting (+8/-6 at this point) it was made clear to me that Gartner's lifecycle is a biased metric from a programmer's perspective. This is something that is part of a paper I'm going to present to management, and management types are part of Gartner's audience. Giving DVCS exposure & enthusiasm (that "could" be deemed as hype, or at least attacked as such), think about the following question when reading this one: "how could I use Gartner's hype cycle to convince management that DVCSs are ready (or ready-enough) for us, and that it is not just hype" Just asking if DVCSs is hype wouldn't be constructive, Gartner's hype cycle is a more objective instrument than just asking that (even if this instrument is regarded as biased). If you know any other instrument please, by all means, mention it. Edit #2: I agree that Gartner's Life Cycle is not for every technology, but I consider it may have generated enough buzz to be considered hype by some, so it maybe deserves to be at least evaluated/pondered as such by using this instrument in order to prove/disprove it to whatever degree. I'm an advocate of DVCS, BTW. I'm doing research for a whitepaper I'm writing in favor of DVCS adoption at company and I stumbled upon the concept of social proof. I want to prove that the social proof of DVCS adoption is not necessarily cargo cult and doing further research I now stumbled upon Gartner's hype cycle which describes technology maturity in 5 phases. My question is: what could be an indicator of the current location of Distributed Version Control Systems (I mean git, mercurial, bazaar, etc. in general) at a particular phase in the hype cycle?... in other (less convoluted) words, would you say that currently expectations of DVCSs are a) starting, b)inflated, c)decreasing (disillusionment), d)increasing (enlightenment) or e)stabilizing (mature) and (more importantly) why? I know it is a hard question and there is subjectivity involved, but I'll grant the answer (and the traditional cookie) to the clearest argument/evidence for a particular phase.

    Read the article

  • Ubuntu and racadm

    - by lmqcn
    I recently purchased a used poweredge 1850 server and it came with a DRAC card. After wiping the HDD and installing ubuntu server 12.04.3 LTS amd64 on it, I am now trying to gain access to the DRAC which I believe is version 4. I have properly configured the DRAC to use it's own IP on my LAN and when I point my browser to the IP address, I am greeted with the DRAC login page (it has the dell logo and everything). However, after trying the credentials of root/calvin, I was denied access. So I think that the previous owners had set their own password. After doing some reading, it appears that I can reset the credentials to the default using racadm config -g cfgUserAdmin -o cfgUserAdminPassword -i 1 newpassword but upon entering the command, I get this error: bash: /usr/sbin/racadm: No such file or directory This holds true even if I run sudo su prior to running the racadm command. If, however, I run sudo racadm config -g cfgUserAdmin -o cfgUserAdminPassword -i 1 newpassword there are no errors. Yet, when I try to log into the DRAC via the web interface using the credentials of root/newpassword I am still not granted access. I installed the dell utilities via the guide at https://wiki.ubuntu.com/HardwareSupportMachinesServersDellNotes. I first tried to install the 64 bit version that is on the dell repositories, but after that was unsuccessful, I just followed the guide verbatim. No errors were produced in either case. I even followed the information at the bottom of the guide by executing sudo pppd /dev/ttyS1 1382400 crtscts noipdefault noauth lock persist connect 'chat -v "" CLIENT CLIENTSERVER "\\c"' but obviously, replacing the /dev/ttyS1 with the correct information for my system. ls -l /usr/sbin/ | grep racadm yields -rwxr-xr-x 1 root root 87930 Sep 16 04:03 racadm I have tried these credentials after each attempt of changing the password: root/calvin root/newpassword admin/calvin admin/newpassword All have been unsuccessful. What is the next course of action that I should take?

    Read the article

  • Hide email adress with JavaScript

    - by Martin Aleksander
    I read somewhere that hiding email address behind JavaScript code, could reduce spam bots harvesting the email address. <script language="javascript" type="text/javascript"> var a = "Red"; var t = "no"; var doc = document; var b = "ITpro"; var ad = a; ad += "@"; ad += b; ad += "."; ad += t; var mt = "ma"; mt += "il"; mt += "to"; var text = ""; if (text == null || text.length == 0) text = ad; doc.write("<"+"a hr"+"ef=\""+mt+":"+ad+"\">"+text+"</"+"a>"); </script> This will not display the actual email-address in the sourcecode of the page, but it will display and work like a normal link for human users. Is it any point of doing this? Will it reduce spam bots, or is it just nonsense that might slow down performance of the page because of the JavaScript?

    Read the article

  • How do I configure an ordinary TV remote control to work with lirc on Linux?

    - by Allan Lewis
    I am running MythTV on Ubuntu 9.10 and I would like to use a TV remote to control it. I know that lirc needs a configuration file for the remote, but none of my remotes is in the official database. If I point a remote at the receiver on my TV card (a Pinnacle PCTV "Solo", model 72e) and press a button, dmesg logs the code generated by the remote, so I assume I just have to make a config file with a list of commands assigned to these codes. I've read a few how-tos but I still don't understand exactly how to create the config file. Some of the guides I've read refer to IR receivers on TV cards working at a "higher level of abstraction", which I take to mean that they decode the signal and provide a code, like the ones I can see in dmesg, rather than just giving raw data, but none of them explain where to go from there! Any help would be greatly appreciated!

    Read the article

  • Chrome Java Plugin not working after moving Java installation directory

    - by Florian Peschka
    I recently moved all my Java installations from C:\Program Files\Java to D:\Java, as my C:\ partition only had a few MB of free space left and windows kept on bugging me about it. I also edited every registry entry in HKEY_LOCAL_MACHINE\SOFTWARE\JavaSoft that pointed to the old directory to now point to the new one. Then changed the PATH variables and removed the old files, did a restart, and all Java applications work just fine. Yet, Chrome is the only application that seems to have a problem with what I did, as the Java Plugin isn't recognized any more. Are there any registry entires I may have missed? Maybe in the Chrome registry? How can I make this work again?

    Read the article

  • Timeout Considerations for Solicit Response

    - by Michael Stephenson
    Background One of the clients I work with had been experiencing some issues for a while surrounding web service timeouts.  It's been a little challenging to work through the problems due to limitations in the diagnostic information available from one of the applications, but I learned some interesting things while troubleshooting the problem which don't seem to have been discussed much in the community so I thought I'd share my findings. In the scenario we have BizTalk trying to make calls to a .net web service which was exposed as a WSE 2 endpoint.  In the process BizTalk will try to make a large number of concurrent web service calls to the application, and the backend application has more than enough infrastructure and capability to handle the load. We have configured the <ConnectionManagement> section of the BizTalk configuration file to support up to 100 concurrent connections from each of our 2 BizTalk send servers to the web servers of the application. The problem we were facing was that the BizTalk side was reporting a significant number of timeouts when calling the web service.   One of the biggest issues was the challenge of being able to correlate a message from BizTalk to the IIS log in the .net application and the custom logs in the application especially when there was a fairly large number of servers hosting the web services.  However the key moment came when we were able to identify a specific call which had taken 40 seconds to execute on the server (yes a long time I know but that's a different story!).  Anyway we were able to identify that this had timed out on the BizTalk side.  Based on the normal 2 minute timeout we knew something unexpected was going on. From here I decided to do some experimentation and I wanted to start outside of BizTalk because my hunch was this was not a BizTalk behaviour but something which was being highlighted by BizTalk because of our large load.     Server-side - Sample Web Service To begin with I created a sample web service.  Nothing special just a vanilla asmx web service hosted in IIS6 on Windows 2003 Standard Edition.  The web service is just a hello world style web service as shown in the below picture.  The only key feature is that the server side web method has a 30 second sleep in it and will trace out some information before and after the thread is set to sleep.      In the configuration for this web service there again is nothing special it's pretty much the most plain simple web service you could build. Client-Side To begin looking at what was happening with our example I created a number of different ways to consume the web service. SoapHttpClientProtocol Example I created a small application which would use a normal proxy generated to call the web service.  It would iterate around a loop and make calls using the begin/end methods so I can do this asynchronously.  I would do a loop of 20 calls with the ConnectionManager configuration section supporting only 5 concurrent connections to the server.     <connectionManagement> <remove address="*"/> <add address = "*" maxconnection = "12" /> <add address = "http://<ServerName>" maxconnection = "5" />                         </connectionManagement> </system.net>     The below picture shows an example of the service calling code, key points are: I have configured the timeout of 40 seconds for the proxy I am using the asynchronous methods on the proxy to call the web service         The Test I would run the client and execute 21 calls to the web service.   The Results  Below is the client side trace showing what's happening on the client. In the below diagram is the web service side trace showing what's happening on the server Some observations on the results are: All of the calls were successful from the clients perspective You could see the next call starting on the server as soon as the previous one had completed Calls took significantly longer than 40 seconds from the start of our call to the return. In fact call 20 took 2 minutes and 30 seconds from the perspective of my code to execute even though I had set the timeout to 40 seconds     WSE 2 Sample In the second example I used the exact same code to call the web service again with a single exception that I modified the web service proxy to derive from WebServiceClient protocol which is part of WSE 2 (using SP3).  The below picture shows the basic code and the key points are: I have configured the timeout of 40 seconds for the proxy I am using the asynchronous methods on the proxy to call the web service        The Test This test would execute 21 calls from the client to the web service.   The Results  The below trace is from the client side: The below trace is from the server side:   Some observations on the trace results for this scenario are: With call 4 if you look at the server side trace it did not start executing on the server for a number of seconds after the other 4 initial calls which were accepted by the server. I re-ran the test and this happened a couple of times and not on most others so at this point I'm just putting this down to something unexpected happening on the development machine and we will leave this observation out of scope of this article. You can see that the client side trace statement executed almost immediately in all cases All calls after the initial few calls would timeout On the client side the calls that did timeout; timed out in a longer duration than the 40 seconds we set as the timeout You can see that as calls were completing on the server the next calls were starting to come through The calls that timed out on the client did actually connect to the server and their server side execution completed successfully     Elaboration on the findings Based on the above observations I have drawn the below sequence diagram to illustrate conceptually what is happening.  Everything except the final web service object is on the client side of the call. In the diagram below I've put two notes on the Web Service Proxy to show the two different places where the different base classes seem to start their timeout counters. From the earlier samples we can work out that the timeout counter for the WSE web service proxy starts before the one for the SoapHttpClientProtocol proxy and the WSE one includes the time to get a connection from the pool; whereas the Soap proxy timeout just covers the method execution. One interesting observation is if we rerun the above sample and increase the number of calls from 21 to 100,000 then for the WSE sample we will see a similar pattern where everything after the first few calls will timeout on the client as soon as it makes a connection to the server whereas the soap proxy will happily plug away and process all of the calls without a single timeout. I have actually set the sample running overnight and this did happen. At this point you are probably thinking the same thoughts I was at the time about the differences in behaviour and which is right and why are they different? I'm not sure there is a definitive answer to this in the documentation, or at least not that I could find! I think you just have to consider that they are different and they could have different effects depending on your messaging solution. In lots of situations this is just not an issue as your concurrent requests doesn't get to the situation where you end up throttling the web service calls on the client side, however this is definitely more common with an integration broker such as BizTalk where you often have high throughput requirements.  Some of the considerations you should make Based on this behaviour you should be aware of the following: In a .net application if you are making lots of concurrent web service calls from an application in an asynchronous manner your user may thing they are experiencing poor performance but you think your web service is working well. The problem could be that the client will have a default of 2 connections to remote servers so you should bear this in mind When you are developing a BizTalk solution or a .net solution with the WSE 2 stack you may experience timeouts under load and throttling the number of connections using the max connections element in the configuration file will not help you For an application using WSE2 or SoapHttpClientProtocol an expired timeout will not throw an error until after a connection to the server has been made so you should consider this in your transaction and durability patterns     Our Work Around In the short term for our specific scenario we know that we can handle this by just increasing our timeout value.  There is only a specific small window when we get lots of concurrent traffic that causes this scenario so we should be able to increase the timeout to take into consideration the additional client side wait, and on the odd occasion where we do get a timeout the BizTalk send port retry will handle this. What was causing our original problem was that for that short window we were getting a lot of retries which significantly increased the load on our send servers and highlighted the issue.  Longer Term Solution As a longer term solution this really gives us more ammunition to argue a migration to WCF. The application we are calling has some factors which limit the protocols we can use but with WCF we would have more control on the various timeout options because in WCF you can configure specific parts of the timeout. Summary I've had this blog post on my to do list for ages but hopefully it will be useful to some people to just understand this behaviour and to possibly help you with some performance issues you may have. I do not believe there is too much in the way of documentation particularly around WSE2 and ASMX in this area so again another bit of ammunition for migrating to WCF. I'll try to do a follow up post with the sample for WCF to show how this changes things.

    Read the article

< Previous Page | 528 529 530 531 532 533 534 535 536 537 538 539  | Next Page >