Search Results

Search found 11930 results on 478 pages for 'shared machines'.

Page 83/478 | < Previous Page | 79 80 81 82 83 84 85 86 87 88 89 90  | Next Page >

• Add users in Windows machine without AD

  - by guillem

  I have several development machines where I am the administrator. We are using AD in my organization but is maintained by and offshore IT group any request takes a long time. We are currently granting access to developers on development machines manually so it's a bit annoying to maintain although at least it's fast. We have also a lot of external consultants that need to use those machines for some time. Is there any tool or method to maintain a set of users synced on those machines without the need add them to an AD group?

  Read the article

• Is it possible to view the contents of an underlying NFS mount without unmounting the NFS content?

  - by Brent

  I have a shared directory on a server - let's call it /home/shared - which is mounted with content from another server via nfs. When it is unmounted /home/shared is supposed to be empty - however, running du -x on the directory indicates that it is not empty. I cannot unmount the NFS content to inspect the mount point, since it is in use by others. Is there any way that I can view/edit the contents of the actual mount point (not the NFS content) while leaving the NFS content mounted for others to use?

  Read the article

• Per Sender Traffic Limit on Cisco 6500

  - by user71557

  Hi All, I have a 6509 with ~1000 user machines in different vlans, I want to allow 10 server machines to send as much as they can/want but to limit all client machines from all subnets to have a sedning rate limit of 1mbps with no receiving limitation. It is worth noting that all my ip addresses are assigned using a DCHP server and there are 1000 of them so I can not write ACLs for every address seperately. Can any one provide some kind of help please?

  Read the article

• Should Production Windows Web Servers (IIS & SQL) be in a domain?

  - by tlianza

  We have a few web servers and a few database servers. To date, they've been standalone machines that are not part of a domain. The web servers don't talk to each other, and the web servers talk to the database servers via SQL Auth. My concern with putting the machines in a domain together were added complexity - it's one more "thing" running, and doing "things" that could go wrong. risk - if a domain controller fails, am I now putting other machines at risk? However, in certain scenarios it does seem convenient for them to be on a domain, sharing credentials. For example, if I want to give the "services" control on one machine access to another machine (because Remote Desktop craps out) I need to go in and assign privileges on multiple machines - something that I believe Active Directory and Domain Accounts set to simplify. My question: I'm sure there are things I'm not considering here. Is there a best practice?

  Read the article

• Folder access per user

  - by user137670

  I have sbs 2003 r2. I have a shared folder (s-drive) for all shared info for everyone. when user is on shared folder, you see size of folder 230G. I have one user that only sees 1g when on shared folder. I have pcs using XP pro. Have check quota and they say no quota limit checked. I had user use a different pc and still same result. With this I looked at server and users profile and compared with user that did not have problem. could not see anything different. what did I miss in some option or do I have to rebuild user? I have tried google with different terms but have not gotten any good clues

  Read the article

• Confusion for mime files: magic, magic.mgc, magic.mime

  - by Florence Foo

  I'm using Ubuntu. I'm trying to use ruby gem 'shared-mime-info' for an application I'm writing. I understand that magic.mgc is a compiled version of magic file which has magic number definitions for the different file types. BUT I don't understand why is it /usr/share/mime/magic is in binary format instead of just normal text file with each parameters separated by white space like everywhere else I'm finding on the internet when it's referencing this file? The /usr/share/mime/magic has the word 'MIME-Magic' at the beginning of the file and prioritize the rest of the stuff like. So it doesn't look like magic.mgc at all. [100:application/vnd.scribus] >1=^@^KSCRIBUSUTF8 [90:application/vnd.stardivision.writer] >2089=^@ shared-mime-info seems to want a magic file in the binary non compiled format as above and I wanted to add definition for DOCX but how does one update or generate this file without using a hex editor? There is a reference to the magic file I found at: http://standards.freedesktop.org/shared-mime-info-spec/shared-mime-info-spec-latest.html And it mention this file is updated with update-mime-database but what if I just want to add some new entry to it. hex editor? Anyway I ended up using hexer to make a new magic file in ~/.local/share/mime/ with only the entry I wanted to add and the MIME-Magic header. Seems to work (assuming I will ever deal with docx for now). 00000000: 4d 49 4d 45 2d 4d 61 67 69 63 00 0a 5b 36 30 3a MIME-Magic..[60: 00000010: 61 70 70 6c 69 63 61 74 69 6f 6e 2f 76 6e 64 2e application/vnd. 00000020: 6f 70 65 6e 78 6d 6c 66 6f 72 6d 61 74 73 2d 6f openxmlformats-o 00000030: 66 66 69 63 65 64 6f 63 75 6d 65 6e 74 2e 77 6f fficedocument.wo 00000040: 72 64 70 72 6f 63 65 73 73 69 6e 67 6d 6c 2e 64 rdprocessingml.d 00000050: 6f 63 75 6d 65 6e 74 5d 0a 3e 30 3d 00 08 50 4b ocument].>0=..PK 00000060: 03 04 14 00 06 00 0a -- -- -- -- -- -- -- -- -- .......---------

  Read the article

• How many different servers are needed to keep a website running with no downtime? [closed]

  - by Mason Wheeler

  Machines go down. It's a fact of life. They may need to be rebooted for some reason, or they may have a hardware failure, or a power outage. So if I wanted to deploy a website with a server backed by a SQL database, putting the whole thing on one server wouldn't be good enough. It obviously needs at least two servers, so that if one goes down, the other can pick up the slack until the first comes back up. Of course, if I have the server software on two machines, either one of which could go down, I can't place the database on either of those two machines, because it could go down. So the database needs its own server. But that server can go down, so I need a backup database server and some sort of replication system to keep it in sync so the main can fail-over to it. So far, that's a bare minimum of 4 machines to keep one website running with a reasonable chance of no downtime. (Assuming no catastrophic events take place that take down both front-end servers at once or both DB servers at once, and no hacks, DDOS attacks, etc. Am I missing any other factors, or should I consider 4 servers to be the minimum for running a website with a goal of continuing operation without downtime even when a server goes down?

  Read the article

• Windows Azure Evolution - Web Sites (aka Antares) Part 1

  - by Shaun

  This is the 3rd post of my Windows Azure Evolution series, focus on the new features and enhancement which was alone with the Windows Azure Platform Upgrade June 2012, announced at the MEET Windows Azure event on 7th June. In the first post I introduced the new preview developer portal and how to works for the existing features such as cloud services, storages and SQL databases. In the second one I talked about the Windows Azure .NET SDK 1.7 on the latest Visual Studio 2012 RC on Windows 8. From this one I will begin to introduce some new features. Now let’s have a look on the first one of them, Windows Azure Web Sites.   Overview Windows Azure Web Sites (WAWS), as known as Antares, was a new feature still in preview stage in this upgrade. It allows people to quickly and easily deploy websites to a highly scalable cloud environment, uses the languages and open source apps of the choice then deploy such as FTP, Git and TFS. It also can be integrated with Windows Azure services like SQL Database, Caching, CDN and Storage easily. After read its introduction we may have a question: since we can deploy a website from both cloud service web role and web site, what’s the different between them? So, let’s have a quick compare.   CLOUD SERVICE WEB SITE OS Windows Server Windows Server Virtualization Windows Azure Virtual Machine Windows Azure Virtual Machine Host IIS IIS Platform ASP.NET WebForm, ASP.NET MVC, WCF ASP.NET WebForm, ASP.NET MVC, PHP Language C#, VB.NET C#, VB.NET, PHP Database SQL Database SQL Database, MySQL Architecture Multi layered, background worker, message queuing, etc.. Simple website with backend database. VS Project Windows Azure Cloud Service ASP.NET Web Form, ASP.NET MVC, etc.. Out-of-box Gallery (none) Drupal, DotNetNuke, WordPress, etc.. Deployment Package upload, Visual Studio publish FTP, Git, TFS, WebMatrix Compute Mode Dedicate VM Shared Across VMs, Dedicate VM Scale Scale up, scale out Scale up, scale out As you can see, there are many difference between the cloud service and web site, but the main point is that, the cloud service focus on those complex architecture web application. For example, if you want to build a website with frontend layer, middle business layer and data access layer, with some background worker process connected through the message queue, then you should better use cloud service, since it provides full control of your code and application. But if you just want to build a personal blog or a  business portal, then you can use the web site. Since the web site have many galleries, you can create them even without any coding and configuration. David Pallmann have an awesome figure explains the benefits between the could service, web site and virtual machine.   Create a Personal Blog in Web Site from Gallery As I mentioned above, one of the big feature in WAWS is to build a website from an existing gallery, which means we don’t need to coding and configure. What we need to do is open the windows azure developer portal and click the NEW button, select WEB SITE and FROM GALLERY. In the popping up windows there are many websites we can choose to use. For example, for personal blog there are Orchard CMS, WordPress; for CMS there are DotNetNuke, Drupal 7, mojoPortal. Let’s select WordPress and click the next button. The next step is to configure the web site. We will need to specify the DNS name and select the subscription and region. Since the WordPress uses MySQL as its backend database, we also need to create a MySQL database as well. Windows Azure Web Sites utilize ClearDB to host the MySQL databases. You cannot create a MySQL database directly from SQL Databases section. Finally, since we selected to create a new MySQL database we need to specify the database name and region in the last step. Also we need to accept the ClearDB’s terms as well. Then windows azure platform will download the WordPress codes and deploy the MySQL database and website. Then it will be ready to use. Select the website and click the BROWSE button, the WordPress administration page will be shown. After configured the WordPress here is my personal web blog on the cloud. It took me no more than 10 minutes to establish without any coding.   Monitor, Configure, Scale and Linked Resources Let’s click into the website I had just created in the portal and have a look on what we can do. In the website details page where are five sections. - Dashboard The overall information about this website, such as the basic usage status, public URL, compute mode, FTP address, subscription and links that we can specify the deployment credentials, TFS and Git publish setting, etc.. - Monitor Some status information such as the CPU usage, memory usage etc., errors, etc.. We can add more metrics by clicking the ADD METRICS button and the bottom as well. - Configure Here we can set the configurations of our website such as the .NET and PHP runtime version, diagnostics settings, application settings and the IIS default documents. - Scale This is something interesting. In WAWS there are two compute mode or called web site mode. One is “shared”, which means our website will be shared with other web sites in a group of windows azure virtual machines. Each web site have its own process (w3wp.exe) with some sandbox technology to isolate from others. When we need to scaling-out our web site in shared mode, we actually increased the working process count. Hence in shared mode we cannot specify the virtual machine size since they are shared across all web sites. This is a little bit different than the scaling mode of the cloud service (hosted service web role and worker role). The other mode called “dedicate”, which means our web site will use the whole windows azure virtual machine. This is the same hosting behavior as cloud service web role. In web role it will be deployed on the virtual machines we specified and all of them are only used by us. In web sites dedicate mode, it’s the same. In this mode when we scaling-out our web site we will use more virtual machines, and each of them will only host our own website. And we can specify the virtual machine size in this mode. In the developer portal we can select which mode we are using from the scale section. In shared mode we can only specify the instance count, but in dedicate mode we can specify the instance size as well as the instance count. - Linked Resource The MySQL database created alone with the creation of our WordPress web site is a linked resource. We can add more linked resources in this section.   Pricing For the web site itself, since this feature is in preview period if you are using shared mode, then you will get free up to 10 web sites. But if you are using dedicate mode, the price would be the virtual machines you are using. For example, if you are using dedicate and configured two middle size virtual machines then you will pay $230.40 per month. If there is SQL Database linked to your web site then they will be charged separately based on the Pay-As-You-Go price. For example a 1GB web edition database costs $9.99 per month. And the bandwidth will be charged as well. For example 10GB outbound data transfer costs $1.20 per month. For more information about the pricing please have a look at the windows azure pricing page.   Summary Windows Azure Web Sites gives us easier and quicker way to create, develop and deploy website to window azure platform. Comparing with the cloud service web role, the WAWS have many out-of-box gallery we can use directly. So if you just want to build a blog, CMS or business portal you don’t need to learn ASP.NET, you don’t need to learn how to configure DotNetNuke, you don’t need to learn how to prepare PHP and MySQL. By using WAWS gallery you can establish a website within 10 minutes without any lines of code. But in some cases we do need to code by ourselves. We may need to tweak the layout of our pages, or we may have a traditional ASP.NET or PHP web application which needed to migrated to the cloud. Besides the gallery WAWS also provides many features to download, upload code. It also provides the feature to integrate with some version control services such as TFS and Git. And it also provides the deploy approaches through FTP and Web Deploy. In the next post I will demonstrate how to use WebMatrix to download and modify the website, and how to use TFS and Git to deploy automatically one our code changes committed.   Hope this helps, Shaun All documents and related graphics, codes are provided "AS IS" without warranty of any kind. Copyright © Shaun Ziyan Xu. This work is licensed under the Creative Commons License.

  Read the article

• White Paper on Analysis Services Tabular Large-scale Solution #ssas #tabular

  - by Marco Russo (SQLBI)

  Since the first beta of Analysis Services 2012, I worked with many companies designing and implementing solutions based on Analysis Services Tabular. I am glad that Microsoft published a white paper about a case-study using one of these scenarios: An Analysis Services Case Study: Using Tabular Models in a Large-scale Commercial Solution. Alberto Ferrari is the author of the white paper and many people contributed to it. The final result is a very technical document based on a case study, which provides a level of detail that I don’t see often in other case studies (which are usually more marketing-oriented). This white paper has the following structure: Requirements (data model, capacity planning, client tool) Options considered (SQL Server Columnstore Indexes, SSAS Multidimensional, SSAS Tabular) Data Model optimizations (memory compression, query performance, scalability) Partitioning and Processing strategy for near real-time latency Hardware selection (NUMA analysis, Azure VM tests) Scalability tests (estimation of maximum users per node) If you are in charge of evaluating Tabular as analytical engine, or if you have to design your solution based on Tabular, this white paper is a must read. But if you just want to increase your knowledge of Analysis Services, you will find a lot of useful technical information. That said, my favorite quote of the document is the following one, funny but true: […] After several trials, the clear winner was a video gaming machine that one guy on the team used at home. That computer outperformed any available server, running twice as fast as the server-class machines we had in house. At that point, it was clear that the criteria for choosing the server would have to be expanded a bit, simply because it would have been impossible to convince the boss to build a cluster of gaming machines and trust it to serve our customers.  But, honestly, if a business has the flexibility to buy gaming machines (assuming the machines can handle capacity) – do this. Owen Graupman, inContact I want to write a longer discussion about how companies are adopting Tabular in scenarios where it is the hidden engine of a more complex solution (and not the classical “BI system”), because it is more frequent than you might expect (and has several advantages over many alternative approaches).

  Read the article

• Hosting and domain registrations for multiple clients under a single hosting account of mine?

  - by letseatfood

  I am finally getting regular work designing, developing, and deploying websites for small businesses and individuals. So far the websites utilize single-user content management systems, so the websites create, as far as I know, minimal load on the shared servers. I have always required that each of my clients purchase annual shared hosting at Dreamhost. For domain registration, I ask that they register with Dreamhost, but some already have a registered domain elsewhere and this is fine with me. I do this so the billing issues are the client's responsibility, not mine. My question is: Since I can register unlimited domains and connect them to my one shared hosting account at Dreamhost, should I not be requiring clients to individually pay for shared hosting and a domain? Should I actually be paying for one hosting account and then hosting all of my client's websites on that account? As I said before, I currently have each client buy their own hosting, because I feel that, for example, if there is high traffic to their site, there would be less a chance of the site going down than if their site was hosted with many others on one account. I am famous for being long-winded, please let me know if I can clarify at all. Thanks!

  Read the article

• ASP.NET design not SOLID

  - by w0051977

  SOLID principles are described here: http://en.wikipedia.org/wiki/SOLID_%28object-oriented_design%29 I am developing a large ASP.NET app. The previous developer created a few very large classes each with lots of different purposes. It is very difficult to maintain and extend. The classes are deployed to the web server along with the code behind files etc. I want to share a small amount of the app with another application. I am considering moving all of the classes of the ASP.NET web app to a DLL, so the small subset of functionality can be shared. I realise it would be better to only share the classes which contain code to be shared but because of the dependencies this is proving to be very difficult e.g. class A contains code that should be shared, however class A contains references to classes B, C, D, E, F, G etc, so class A cannot be shared on its own. I am planning to refactor the code in the future. As a temporary solution I am planning to convert all the classes into a single class library. Is this a bad idea and if so, is there an alternative? as I don't have time to refactor at the moment.

  Read the article

• Hosting and domain registrations for multiple clients

  - by letseatfood

  I am finally getting regular work desiging, developing, and deploying websites for small businesses and individuals. So far the websites utilize single-user content management systems, so the websites create, as far as I know, minimal load on the shared servers. I have always required that each of my clients purchase annual shared hosting at Dreamhost. For domain registration, I ask that they register with Dreamhost, but some already have a registered domain elsewhere and this is fine with me. I do this so the billing issues are the client's responsibility, not mine. My question is: Since I can register unlimited domains and connect them to my one shared hosting account at Dreamhost, should I not be requiring clients to individually pay for shared hosting and a domain? Should I actually be paying for one hosting account and then hosting all of my client's websites on that account? As I said before, I currently have each client buy their own hosting, because I feel that, for example, if there is high traffic to their site, there would be less a chance of the site going down than if their site was hosted with many others on one account. I am famous for being long-winded, please let me know if I can clarify at all. Thanks!

  Read the article

• Errors when installing updates

  - by user71613

  I am getting the following errors when installing updates. They started to appear after I upgraded my system to 12.04. Errors were encountered while processing: samba-common samba-common-bin samba grub-pc grub-gfxpayload-lists Setting up samba-common (2:3.6.3-2ubuntu2.2) ... perl: error while loading shared libraries: libperl.so.5.12: cannot open shared object file: No such file or directory dpkg: error processing samba-common (--configure): subprocess installed post-installation script returned error exit status 127 dpkg: dependency problems prevent configuration of samba-common-bin: samba-common-bin depends on samba-common (>= 2:3.4.0~pre1-2); however: Package samba-common is not configured yet. dpkg: error processing samba-common-bin (--configure): dependency problems - leaving unconfigured dpkg: dependency problems prevent configuration of samba: samba depends on samba-common (= 2:3.6.3-2ubuntu2.2); however: Package samba-common is not configured yet. samba depends on samba-common-bin; however: Package samba-common-bin is not configured yet. dpkg: error processing samba (--configure): dependency problems - leaving unconfigured Setting up grub-gfxpayload-lists (0.6) ... Setting up grub-pc (1.99-21ubuntu3.1) ... perl: error while loading shared libraries: libperl.so.5.12: cannot open shared object file: No such file or directory dpkg: error processing grub-pc (--configure): subprocess installed post-installation script returned error exit status 127 Any ideas how to fix this?

  Read the article

• DLL Config in Mono

  - by nubela

  Hi, I'm trying to pick up Svn.NET (http://www.pumacode.org/projects/svndotnet/) library for use in my Mono project. I tried compiling its mockapp - svnmockapp project (http://www.pumacode.org/projects/svndotnet/browser/trunk/SvnMockApp) , I am able to get the references right and get it compiled right. I understand that it references 2 other modules libapr (libapr-1.so.0) and svn_client (libsvn_client-1.so.0) , by which I've created PumaCode.SvnDotNet.dll.config in /bin/Debug . That is all I've done to tried to try to get the mockapp at least outputting something to show that it is indeed interfacing SVN. Nevertheless, it is not working. Commands that are entered that doesn't interface SVN works fine: $ mono SvnTest.exe -usage Usage: SvnTest <subcommand> [options] Short Options: ~?.V Subcommands: add, checkout[co], status[st], update[up] For help on subcommands, use the -?/--help subcommand option. Commands that tries to access SVN throws an exception: $ mono SvnTest.exe st An exception was thrown by the type initializer for PumaCode.SvnDotNet.AprSharp.Apr Unhandled Exception: System.Reflection.TargetInvocationException: Exception has been thrown by the target of an invocation. ---> System.NullReferenceException: Object reference not set to an instance of an object at PumaCode.SvnDotNet.SubversionSharp.SvnMockApp.CmdBase.Run (PumaCode.SvnDotNet.SubversionSharp.SvnMockApp.SubCommand sc, System.String[] args) [0x00000] at (wrapper managed-to-native) System.Reflection.MonoMethod:InternalInvoke (object,object[],System.Exception&) at System.Reflection.MonoMethod.Invoke (System.Object obj, BindingFlags invokeAttr, System.Reflection.Binder binder, System.Object[] parameters, System.Globalization.CultureInfo culture) [0x00000] --- End of inner exception stack trace --- at System.Reflection.MonoMethod.Invoke (System.Object obj, BindingFlags invokeAttr, System.Reflection.Binder binder, System.Object[] parameters, System.Globalization.CultureInfo culture) [0x00000] at System.Reflection.MethodBase.Invoke (System.Object obj, System.Object[] parameters) [0x00000] at PumaCode.SvnDotNet.SubversionSharp.SvnMockApp.Application.Run (System.String[] args) [0x00000] at PumaCode.SvnDotNet.SubversionSharp.SvnMockApp.Application.Main (System.String[] args) [0x00000] Using MONO_DEBUG_LEVEL="debug", we get the following log pasted here. At the tail end of the log, we see this: . . . Mono-INFO: DllImport attempting to load: 'libapr-1'. Mono-INFO: DllImport loading location: 'libapr-1.so'. Mono-INFO: DllImport error loading library: 'libapr-1.so: cannot open shared object file: No such file or directory'. Mono-INFO: DllImport loading library: './libapr-1.so'. Mono-INFO: DllImport error loading library './libapr-1.so: cannot open shared object file: No such file or directory'. Mono-INFO: DllImport loading: 'libapr-1'. Mono-INFO: DllImport error loading library 'libapr-1: cannot open shared object file: No such file or directory'. Mono-INFO: DllImport attempting to load: 'libapr-1'. Mono-INFO: DllImport loading location: 'libapr-1.so'. Mono-INFO: DllImport error loading library: 'libapr-1.so: cannot open shared object file: No such file or directory'. Mono-INFO: DllImport loading library: './libapr-1.so'. Mono-INFO: DllImport error loading library './libapr-1.so: cannot open shared object file: No such file or directory'. Mono-INFO: DllImport loading: 'libapr-1'. Mono-INFO: DllImport error loading library 'libapr-1: cannot open shared object file: No such file or directory'. Mono-INFO: DllImport attempting to load: 'libapr-1'. Mono-INFO: DllImport loading location: 'libapr-1.so'. Mono-INFO: DllImport error loading library: 'libapr-1.so: cannot open shared object file: No such file or directory'. Mono-INFO: DllImport loading library: './libapr-1.so'. Mono-INFO: DllImport error loading library './libapr-1.so: cannot open shared object file: No such file or directory'. Mono-INFO: DllImport loading: 'libapr-1'. Mono-INFO: DllImport error loading library 'libapr-1: cannot open shared object file: No such file or directory'. An exception was thrown by the type initializer for PumaCode.SvnDotNet.AprSharp.Apr I've tried to symlink the appropriate modules in the directory where SvnTest.exe exists, but this still persist. How can I fix this? Did I place the PumaCode.SvnDotNet.dll.config in the wrong folder? (I placed it at /bin/Debug and also tried /bin) What can I do to remedy this? Thank you for your kind help! Much appreciated! Heres the config file: (PumaCode.SvnDotNet.dll.config) <configuration> <dllmap dll="libapr" target="/usr/lib/libapr-1.so.0"/> <dllmap dll="svn_client-1" target="/usr/lib/libsvn_client-1.so.0"/> </configuration>

  Read the article

• How to handle failure to release a resource which is contained in a smart pointer?

  - by cj

  How should an error during resource deallocation be handled, when the object representing the resource is contained in a shared pointer? Smart pointers are a useful tool to manage resources safely. Examples of such resources are memory, disk files, database connections, or network connections. // open a connection to the local HTTP port boost::shared_ptr<Socket> socket = Socket::connect("localhost:80"); In a typical scenario, the class encapsulating the resource should be noncopyable and polymorphic. A good way to support this is to provide a factory method returning a shared pointer, and declare all constructors non-public. The shared pointers can now be copied from and assigned to freely. The object is automatically destroyed when no reference to it remains, and the destructor then releases the resource. /** A TCP/IP connection. */ class Socket { public: static boost::shared_ptr<Socket> connect(const std::string& address); virtual ~Socket(); protected: Socket(const std::string& address); private: // not implemented Socket(const Socket&); Socket& operator=(const Socket&); }; But there is a problem with this approach. The destructor must not throw, so a failure to release the resource will remain undetected. A common way out of this problem is to add a public method to release the resource. class Socket { public: virtual void close(); // may throw // ... }; Unfortunately, this approach introduces another problem: Our objects may now contain resources which have already been released. This complicates the implementation of the resource class. Even worse, it makes it possible for clients of the class to use it incorrectly. The following example may seem far-fetched, but it is a common pitfall in multi-threaded code. socket->close(); // ... size_t nread = socket->read(&buffer[0], buffer.size()); // wrong use! Either we ensure that the resource is not released before the object is destroyed, thereby losing any way to deal with a failed resource deallocation. Or we provide a way to release the resource explicitly during the object's lifetime, thereby making it possible to use the resource class incorrectly. There is a way out of this dilemma. But the solution involves using a modified shared pointer class. These modifications are likely to be controversial. Typical shared pointer implementations, such as boost::shared_ptr, require that no exception be thrown when their object's destructor is called. Generally, no destructor should ever throw, so this is a reasonable requirement. These implementations also allow a custom deleter function to be specified, which is called in lieu of the destructor when no reference to the object remains. The no-throw requirement is extended to this custom deleter function. The rationale for this requirement is clear: The shared pointer's destructor must not throw. If the deleter function does not throw, nor will the shared pointer's destructor. However, the same holds for other member functions of the shared pointer which lead to resource deallocation, e.g. reset(): If resource deallocation fails, no exception can be thrown. The solution proposed here is to allow custom deleter functions to throw. This means that the modified shared pointer's destructor must catch exceptions thrown by the deleter function. On the other hand, member functions other than the destructor, e.g. reset(), shall not catch exceptions of the deleter function (and their implementation becomes somewhat more complicated). Here is the original example, using a throwing deleter function: /** A TCP/IP connection. */ class Socket { public: static SharedPtr<Socket> connect(const std::string& address); protected: Socket(const std::string& address); virtual Socket() { } private: struct Deleter; // not implemented Socket(const Socket&); Socket& operator=(const Socket&); }; struct Socket::Deleter { void operator()(Socket* socket) { // Close the connection. If an error occurs, delete the socket // and throw an exception. delete socket; } }; SharedPtr<Socket> Socket::connect(const std::string& address) { return SharedPtr<Socket>(new Socket(address), Deleter()); } We can now use reset() to free the resource explicitly. If there is still a reference to the resource in another thread or another part of the program, calling reset() will only decrement the reference count. If this is the last reference to the resource, the resource is released. If resource deallocation fails, an exception is thrown. SharedPtr<Socket> socket = Socket::connect("localhost:80"); // ... socket.reset();

  Read the article

• Networking in VirtualBox

  - by Fat Bloke

  Networking in VirtualBox is extremely powerful, but can also be a bit daunting, so here's a quick overview of the different ways you can setup networking in VirtualBox, with a few pointers as to which configurations should be used and when. VirtualBox allows you to configure up to 8 virtual NICs (Network Interface Controllers) for each guest vm (although only 4 are exposed in the GUI) and for each of these NICs you can configure: Which virtualized NIC-type is exposed to the Guest. Examples include: Intel PRO/1000 MT Server (82545EM),  AMD PCNet FAST III (Am79C973, the default) or  a Paravirtualized network adapter (virtio-net). How the NIC operates with respect to your Host's physical networking. The main modes are: Network Address Translation (NAT) Bridged networking Internal networking Host-only networking NAT with Port-forwarding The choice of NIC-type comes down to whether the guest has drivers for that NIC.  VirtualBox, suggests a NIC based on the guest OS-type that you specify during creation of the vm, and you rarely need to modify this. But the choice of networking mode depends on how you want to use your vm (client or server) and whether you want other machines on your network to see it. So let's look at each mode in a bit more detail... Network Address Translation (NAT) This is the default mode for new vm's and works great in most situations when the Guest is a "client" type of vm. (i.e. most network connections are outbound). Here's how it works: When the guest OS boots,  it typically uses DHCP to get an IP address. VirtualBox will field this DHCP request and tell the guest OS its assigned IP address and the gateway address for routing outbound connections. In this mode, every vm is assigned the same IP address (10.0.2.15) because each vm thinks they are on their own isolated network. And when they send their traffic via the gateway (10.0.2.2) VirtualBox rewrites the packets to make them appear as though they originated from the Host, rather than the Guest (running inside the Host). This means that the Guest will work even as the Host moves from network to network (e.g. laptop moving between locations), and from wireless to wired connections too. However, how does another computer initiate a connection into a Guest?  e.g. connecting to a web server running in the Guest. This is not (normally) possible using NAT mode as there is no route into the Guest OS. So for vm's running servers we need a different networking mode.... Bridged Networking Bridged Networking is used when you want your vm to be a full network citizen, i.e. to be an equal to your host machine on the network. In this mode, a virtual NIC is "bridged" to a physical NIC on your host, like this: The effect of this is that each VM has access to the physical network in the same way as your host. It can access any service on the network such as external DHCP services, name lookup services, and routing information just as the host does. Logically, the network looks like this: The downside of this mode is that if you run many vm's you can quickly run out of IP addresses or your network administrator gets fed up with you asking for statically assigned IP addresses. Secondly, if your host has multiple physical NICs (e.g. Wireless and Wired) you must reconfigure the bridge when your host jumps networks.  Hmm, so what if you want to run servers in vm's but don't want to involve your network administrator? Maybe one of the next 2 modes is for you... Internal Networking When you configure one or more vm's to sit on an Internal network, VirtualBox ensures that all traffic on that network stays within the host and is only visible to vm's on that virtual network. Configuration looks like this: The internal network ( in this example "intnet" ) is a totally isolated network and so is very "quiet". This is good for testing when you need a separate, clean network, and you can create sophisticated internal networks with vm's that provide their own services to the internal network. (e.g. Active Directory, DHCP, etc). Note that not even the Host is a member of the internal network, but this mode allows vm's to function even when the Host is not connected to a network (e.g. on a plane). Note that in this mode, VirtualBox provides no "convenience" services such as DHCP, so your machines must be statically configured or one of the vm's needs to provide a DHCP/Name service. Multiple internal networks are possible and you can configure vm's to have multiple NICs to sit across internal and other network modes and thereby provide routes if needed. But all this sounds tricky. What if you want an Internal Network that the host participates on with VirtualBox providing IP addresses to the Guests? Ah, then for this, you might want to consider Host-only Networking... Host-only Networking Host-only Networking is like Internal Networking in that you indicate which network the Guest sits on, in this case, "vboxnet0": All vm's sitting on this "vboxnet0" network will see each other, and additionally, the host can see these vm's too. However, other external machines cannot see Guests on this network, hence the name "Host-only". Logically, the network looks like this: This looks very similar to Internal Networking but the host is now on "vboxnet0" and can provide DHCP services. To configure how a Host-only network behaves, look in the VirtualBox Manager...Preferences...Network dialog: Port-Forwarding with NAT Networking Now you may think that we've provided enough modes here to handle every eventuality but here's just one more... What if you cart around a mobile-demo or dev environment on, say, a laptop and you have one or more vm's that you need other machines to connect into? And you are continually hopping onto different (customer?) networks. In this scenario: NAT - won't work because external machines need to connect in. Bridged - possibly an option, but does your customer want you eating IP addresses and can your software cope with changing networks? Internal - we need the vm(s) to be visible on the network, so this is no good. Host-only - same problem as above, we want external machines to connect in to the vm's. Enter Port-forwarding to save the day! Configure your vm's to use NAT networking; Add Port Forwarding rules; External machines connect to "host":"port number" and connections are forwarded by VirtualBox to the guest:port number specified. For example, if your vm runs a web server on port 80, you could set up rules like this:  ...which reads: "any connections on port 8080 on the Host will be forwarded onto this vm's port 80".  This provides a mobile demo system which won't need re-configuring every time you open your laptop lid. Summary VirtualBox has a very powerful set of options allowing you to set up almost any configuration your heart desires. For more information, check out the VirtualBox User Manual on Virtual Networking. -FB 

  Read the article

• Windows Azure: Backup Services Release, Hyper-V Recovery Manager, VM Enhancements, Enhanced Enterprise Management Support

  - by ScottGu

  This morning we released a huge set of updates to Windows Azure.  These new capabilities include: Backup Services: General Availability of Windows Azure Backup Services Hyper-V Recovery Manager: Public preview of Windows Azure Hyper-V Recovery Manager Virtual Machines: Delete Attached Disks, Availability Set Warnings, SQL AlwaysOn Configuration Active Directory: Securely manage hundreds of SaaS applications Enterprise Management: Use Active Directory to Better Manage Windows Azure Windows Azure SDK 2.2: A massive update of our SDK + Visual Studio tooling support All of these improvements are now available to use immediately.  Below are more details about them. Backup Service: General Availability Release of Windows Azure Backup Today we are releasing Windows Azure Backup Service as a general availability service.  This release is now live in production, backed by an enterprise SLA, supported by Microsoft Support, and is ready to use for production scenarios. Windows Azure Backup is a cloud based backup solution for Windows Server which allows files and folders to be backed up and recovered from the cloud, and provides off-site protection against data loss. The service provides IT administrators and developers with the option to back up and protect critical data in an easily recoverable way from any location with no upfront hardware cost. Windows Azure Backup is built on the Windows Azure platform and uses Windows Azure blob storage for storing customer data. Windows Server uses the downloadable Windows Azure Backup Agent to transfer file and folder data securely and efficiently to the Windows Azure Backup Service. Along with providing cloud backup for Windows Server, Windows Azure Backup Service also provides capability to backup data from System Center Data Protection Manager and Windows Server Essentials, to the cloud. All data is encrypted onsite before it is sent to the cloud, and customers retain and manage the encryption key (meaning the data is stored entirely secured and can’t be decrypted by anyone but yourself). Getting Started To get started with the Windows Azure Backup Service, create a new Backup Vault within the Windows Azure Management Portal.  Click New->Data Services->Recovery Services->Backup Vault to do this: Once the backup vault is created you’ll be presented with a simple tutorial that will help guide you on how to register your Windows Servers with it: Once the servers you want to backup are registered, you can use the appropriate local management interface (such as the Microsoft Management Console snap-in, System Center Data Protection Manager Console, or Windows Server Essentials Dashboard) to configure the scheduled backups and to optionally initiate recoveries. You can follow these tutorials to learn more about how to do this: Tutorial: Schedule Backups Using the Windows Azure Backup Agent This tutorial helps you with setting up a backup schedule for your registered Windows Servers. Additionally, it also explains how to use Windows PowerShell cmdlets to set up a custom backup schedule. Tutorial: Recover Files and Folders Using the Windows Azure Backup Agent This tutorial helps you with recovering data from a backup. Additionally, it also explains how to use Windows PowerShell cmdlets to do the same tasks. Below are some of the key benefits the Windows Azure Backup Service provides: Simple configuration and management. Windows Azure Backup Service integrates with the familiar Windows Server Backup utility in Windows Server, the Data Protection Manager component in System Center and Windows Server Essentials, in order to provide a seamless backup and recovery experience to a local disk, or to the cloud. Block level incremental backups. The Windows Azure Backup Agent performs incremental backups by tracking file and block level changes and only transferring the changed blocks, hence reducing the storage and bandwidth utilization. Different point-in-time versions of the backups use storage efficiently by only storing the changes blocks between these versions. Data compression, encryption and throttling. The Windows Azure Backup Agent ensures that data is compressed and encrypted on the server before being sent to the Windows Azure Backup Service over the network. As a result, the Windows Azure Backup Service only stores encrypted data in the cloud storage. The encryption key is not available to the Windows Azure Backup Service, and as a result the data is never decrypted in the service. Also, users can setup throttling and configure how the Windows Azure Backup service utilizes the network bandwidth when backing up or restoring information. Data integrity is verified in the cloud. In addition to the secure backups, the backed up data is also automatically checked for integrity once the backup is done. As a result, any corruptions which may arise due to data transfer can be easily identified and are fixed automatically. Configurable retention policies for storing data in the cloud. The Windows Azure Backup Service accepts and implements retention policies to recycle backups that exceed the desired retention range, thereby meeting business policies and managing backup costs. Hyper-V Recovery Manager: Now Available in Public Preview I’m excited to also announce the public preview of a new Windows Azure Service – the Windows Azure Hyper-V Recovery Manager (HRM). Windows Azure Hyper-V Recovery Manager helps protect your business critical services by coordinating the replication and recovery of System Center Virtual Machine Manager 2012 SP1 and System Center Virtual Machine Manager 2012 R2 private clouds at a secondary location. With automated protection, asynchronous ongoing replication, and orderly recovery, the Hyper-V Recovery Manager service can help you implement Disaster Recovery and restore important services accurately, consistently, and with minimal downtime. Application data in an Hyper-V Recovery Manager scenarios always travels on your on-premise replication channel. Only metadata (such as names of logical clouds, virtual machines, networks etc.) that is needed for orchestration is sent to Azure. All traffic sent to/from Azure is encrypted. You can begin using Windows Azure Hyper-V Recovery today by clicking New->Data Services->Recovery Services->Hyper-V Recovery Manager within the Windows Azure Management Portal.  You can read more about Windows Azure Hyper-V Recovery Manager in Brad Anderson’s 9-part series, Transform the datacenter. To learn more about setting up Hyper-V Recovery Manager follow our detailed step-by-step guide. Virtual Machines: Delete Attached Disks, Availability Set Warnings, SQL AlwaysOn Today’s Windows Azure release includes a number of nice updates to Windows Azure Virtual Machines.  These improvements include: Ability to Delete both VM Instances + Attached Disks in One Operation Prior to today’s release, when you deleted VMs within Windows Azure we would delete the VM instance – but not delete the drives attached to the VM.  You had to manually delete these yourself from the storage account.  With today’s update we’ve added a convenience option that now allows you to either retain or delete the attached disks when you delete the VM:   We’ve also added the ability to delete a cloud service, its deployments, and its role instances with a single action. This can either be a cloud service that has production and staging deployments with web and worker roles, or a cloud service that contains virtual machines.  To do this, simply select the Cloud Service within the Windows Azure Management Portal and click the “Delete” button: Warnings on Availability Sets with Only One Virtual Machine In Them One of the nice features that Windows Azure Virtual Machines supports is the concept of “Availability Sets”.  An “availability set” allows you to define a tier/role (e.g. webfrontends, databaseservers, etc) that you can map Virtual Machines into – and when you do this Windows Azure separates them across fault domains and ensures that at least one of them is always available during servicing operations.  This enables you to deploy applications in a high availability way. One issue we’ve seen some customers run into is where they define an availability set, but then forget to map more than one VM into it (which defeats the purpose of having an availability set).  With today’s release we now display a warning in the Windows Azure Management Portal if you have only one virtual machine deployed in an availability set to help highlight this: You can learn more about configuring the availability of your virtual machines here. Configuring SQL Server Always On SQL Server Always On is a great feature that you can use with Windows Azure to enable high availability and DR scenarios with SQL Server. Today’s Windows Azure release makes it even easier to configure SQL Server Always On by enabling “Direct Server Return” endpoints to be configured and managed within the Windows Azure Management Portal.  Previously, setting this up required using PowerShell to complete the endpoint configuration.  Starting today you can enable this simply by checking the “Direct Server Return” checkbox: You can learn more about how to use direct server return for SQL Server AlwaysOn availability groups here. Active Directory: Application Access Enhancements This summer we released our initial preview of our Application Access Enhancements for Windows Azure Active Directory.  This service enables you to securely implement single-sign-on (SSO) support against SaaS applications (including Office 365, SalesForce, Workday, Box, Google Apps, GitHub, etc) as well as LOB based applications (including ones built with the new Windows Azure AD support we shipped last week with ASP.NET and VS 2013). Since the initial preview we’ve enhanced our SAML federation capabilities, integrated our new password vaulting system, and shipped multi-factor authentication support. We've also turned on our outbound identity provisioning system and have it working with hundreds of additional SaaS Applications: Earlier this month we published an update on dates and pricing for when the service will be released in general availability form.  In this blog post we announced our intention to release the service in general availability form by the end of the year.  We also announced that the below features would be available in a free tier with it: SSO to every SaaS app we integrate with – Users can Single Sign On to any app we are integrated with at no charge. This includes all the top SAAS Apps and every app in our application gallery whether they use federation or password vaulting. Application access assignment and removal – IT Admins can assign access privileges to web applications to the users in their active directory assuring that every employee has access to the SAAS Apps they need. And when a user leaves the company or changes jobs, the admin can just as easily remove their access privileges assuring data security and minimizing IP loss User provisioning (and de-provisioning) – IT admins will be able to automatically provision users in 3rd party SaaS applications like Box, Salesforce.com, GoToMeeting, DropBox and others. We are working with key partners in the ecosystem to establish these connections, meaning you no longer have to continually update user records in multiple systems. Security and auditing reports – Security is a key priority for us. With the free version of these enhancements you'll get access to our standard set of access reports giving you visibility into which users are using which applications, when they were using them and where they are using them from. In addition, we'll alert you to un-usual usage patterns for instance when a user logs in from multiple locations at the same time. Our Application Access Panel – Users are logging in from every type of devices including Windows, iOS, & Android. Not all of these devices handle authentication in the same manner but the user doesn't care. They need to access their apps from the devices they love. Our Application Access Panel will support the ability for users to access access and launch their apps from any device and anywhere. You can learn more about our plans for application management with Windows Azure Active Directory here.  Try out the preview and start using it today. Enterprise Management: Use Active Directory to Better Manage Windows Azure Windows Azure Active Directory provides the ability to manage your organization in a directory which is hosted entirely in the cloud, or alternatively kept in sync with an on-premises Windows Server Active Directory solution (allowing you to seamlessly integrate with the directory you already have).  With today’s Windows Azure release we are integrating Windows Azure Active Directory even more within the core Windows Azure management experience, and enabling an even richer enterprise security offering.  Specifically: 1) All Windows Azure accounts now have a default Windows Azure Active Directory created for them.  You can create and map any users you want into this directory, and grant administrative rights to manage resources in Windows Azure to these users. 2) You can keep this directory entirely hosted in the cloud – or optionally sync it with your on-premises Windows Server Active Directory.  Both options are free.  The later approach is ideal for companies that wish to use their corporate user identities to sign-in and manage Windows Azure resources.  It also ensures that if an employee leaves an organization, his or her access control rights to the company’s Windows Azure resources are immediately revoked. 3) The Windows Azure Service Management APIs have been updated to support using Windows Azure Active Directory credentials to sign-in and perform management operations.  Prior to today’s release customers had to download and use management certificates (which were not scoped to individual users) to perform management operations.  We still support this management certificate approach (don’t worry – nothing will stop working).  But we think the new Windows Azure Active Directory authentication support enables an even easier and more secure way for customers to manage resources going forward.  4) The Windows Azure SDK 2.2 release (which is also shipping today) includes built-in support for the new Service Management APIs that authenticate with Windows Azure Active Directory, and now allow you to create and manage Windows Azure applications and resources directly within Visual Studio using your Active Directory credentials.  This, combined with updated PowerShell scripts that also support Active Directory, enables an end-to-end enterprise authentication story with Windows Azure. Below are some details on how all of this works: Subscriptions within a Directory As part of today’s update, we have associated all existing Window Azure accounts with a Windows Azure Active Directory (and created one for you if you don’t already have one). When you login to the Windows Azure Management Portal you’ll now see the directory name in the URI of the browser.  For example, in the screen-shot below you can see that I have a “scottgu” directory that my subscriptions are hosted within: Note that you can continue to use Microsoft Accounts (formerly known as Microsoft Live IDs) to sign-into Windows Azure.  These map just fine to a Windows Azure Active Directory – so there is no need to create new usernames that are specific to a directory if you don’t want to.  In the scenario above I’m actually logged in using my @hotmail.com based Microsoft ID which is now mapped to a “scottgu” active directory that was created for me.  By default everything will continue to work just like you used to before. Manage your Directory You can manage an Active Directory (including the one we now create for you by default) by clicking the “Active Directory” tab in the left-hand side of the portal.  This will list all of the directories in your account.  Clicking one the first time will display a getting started page that provides documentation and links to perform common tasks with it: You can use the built-in directory management support within the Windows Azure Management Portal to add/remove/manage users within the directory, enable multi-factor authentication, associate a custom domain (e.g. mycompanyname.com) with the directory, and/or rename the directory to whatever friendly name you want (just click the configure tab to do this).  You can also setup the directory to automatically sync with an on-premises Active Directory using the “Directory Integration” tab. Note that users within a directory by default do not have admin rights to login or manage Windows Azure based resources.  You still need to explicitly grant them co-admin permissions on a subscription for them to login or manage resources in Windows Azure.  You can do this by clicking the Settings tab on the left-hand side of the portal and then by clicking the administrators tab within it. Sign-In Integration within Visual Studio If you install the new Windows Azure SDK 2.2 release, you can now connect to Windows Azure from directly inside Visual Studio without having to download any management certificates.  You can now just right-click on the “Windows Azure” icon within the Server Explorer and choose the “Connect to Windows Azure” context menu option to do so: Doing this will prompt you to enter the email address of the username you wish to sign-in with (make sure this account is a user in your directory with co-admin rights on a subscription): You can use either a Microsoft Account (e.g. Windows Live ID) or an Active Directory based Organizational account as the email.  The dialog will update with an appropriate login prompt depending on which type of email address you enter: Once you sign-in you’ll see the Windows Azure resources that you have permissions to manage show up automatically within the Visual Studio server explorer and be available to start using: No downloading of management certificates required.  All of the authentication was handled using your Windows Azure Active Directory! Manage Subscriptions across Multiple Directories If you have already have multiple directories and multiple subscriptions within your Windows Azure account, we have done our best to create a good default mapping of your subscriptions->directories as part of today’s update.  If you don’t like the default subscription-to-directory mapping we have done you can click the Settings tab in the left-hand navigation of the Windows Azure Management Portal and browse to the Subscriptions tab within it: If you want to map a subscription under a different directory in your account, simply select the subscription from the list, and then click the “Edit Directory” button to choose which directory to map it to.  Mapping a subscription to a different directory takes only seconds and will not cause any of the resources within the subscription to recycle or stop working.  We’ve made the directory->subscription mapping process self-service so that you always have complete control and can map things however you want. Filtering By Directory and Subscription Within the Windows Azure Management Portal you can filter resources in the portal by subscription (allowing you to show/hide different subscriptions).  If you have subscriptions mapped to multiple directory tenants, we also now have a filter drop-down that allows you to filter the subscription list by directory tenant.  This filter is only available if you have multiple subscriptions mapped to multiple directories within your Windows Azure Account:   Windows Azure SDK 2.2 Today we are also releasing a major update of our Windows Azure SDK.  The Windows Azure SDK 2.2 release adds some great new features including: Visual Studio 2013 Support Integrated Windows Azure Sign-In support within Visual Studio Remote Debugging Cloud Services with Visual Studio Firewall Management support within Visual Studio for SQL Databases Visual Studio 2013 RTM VM Images for MSDN Subscribers Windows Azure Management Libraries for .NET Updated Windows Azure PowerShell Cmdlets and ScriptCenter I’ll post a follow-up blog shortly with more details about all of the above. Additional Updates In addition to the above enhancements, today’s release also includes a number of additional improvements: AutoScale: Richer time and date based scheduling support (set different rules on different dates) AutoScale: Ability to Scale to Zero Virtual Machines (very useful for Dev/Test scenarios) AutoScale: Support for time-based scheduling of Mobile Service AutoScale rules Operation Logs: Auditing support for Service Bus management operations Today we also shipped a major update to the Windows Azure SDK – Windows Azure SDK 2.2.  It has so much goodness in it that I have a whole second blog post coming shortly on it! :-) Summary Today’s Windows Azure release enables a bunch of great new scenarios, and enables a much richer enterprise authentication offering. If you don’t already have a Windows Azure account, you can sign-up for a free trial and start using all of the above features today.  Then visit the Windows Azure Developer Center to learn more about how to build apps with it. Hope this helps, Scott P.S. In addition to blogging, I am also now using Twitter for quick updates and to share links. Follow me at: twitter.com/scottgu

  Read the article

• Why lock-free data structures just aren't lock-free enough

  - by Alex.Davies

  Today's post will explore why the current ways to communicate between threads don't scale, and show you a possible way to build scalable parallel programming on top of shared memory. The problem with shared memory Soon, we will have dozens, hundreds and then millions of cores in our computers. It's inevitable, because individual cores just can't get much faster. At some point, that's going to mean that we have to rethink our architecture entirely, as millions of cores can't all access a shared memory space efficiently. But millions of cores are still a long way off, and in the meantime we'll see machines with dozens of cores, struggling with shared memory. Alex's tip: The best way for an application to make use of that increasing parallel power is to use a concurrency model like actors, that deals with synchronisation issues for you. Then, the maintainer of the actors framework can find the most efficient way to coordinate access to shared memory to allow your actors to pass messages to each other efficiently. At the moment, NAct uses the .NET thread pool and a few locks to marshal messages. It works well on dual and quad core machines, but it won't scale to more cores. Every time we use a lock, our core performs an atomic memory operation (eg. CAS) on a cell of memory representing the lock, so it's sure that no other core can possibly have that lock. This is very fast when the lock isn't contended, but we need to notify all the other cores, in case they held the cell of memory in a cache. As the number of cores increases, the total cost of a lock increases linearly. A lot of work has been done on "lock-free" data structures, which avoid locks by using atomic memory operations directly. These give fairly dramatic performance improvements, particularly on systems with a few (2 to 4) cores. The .NET 4 concurrent collections in System.Collections.Concurrent are mostly lock-free. However, lock-free data structures still don't scale indefinitely, because any use of an atomic memory operation still involves every core in the system. A sync-free data structure Some concurrent data structures are possible to write in a completely synchronization-free way, without using any atomic memory operations. One useful example is a single producer, single consumer (SPSC) queue. It's easy to write a sync-free fixed size SPSC queue using a circular buffer*. Slightly trickier is a queue that grows as needed. You can use a linked list to represent the queue, but if you leave the nodes to be garbage collected once you're done with them, the GC will need to involve all the cores in collecting the finished nodes. Instead, I've implemented a proof of concept inspired by this intel article which reuses the nodes by putting them in a second queue to send back to the producer. * In all these cases, you need to use memory barriers correctly, but these are local to a core, so don't have the same scalability problems as atomic memory operations. Performance tests I tried benchmarking my SPSC queue against the .NET ConcurrentQueue, and against a standard Queue protected by locks. In some ways, this isn't a fair comparison, because both of these support multiple producers and multiple consumers, but I'll come to that later. I started on my dual-core laptop, running a simple test that had one thread producing 64 bit integers, and another consuming them, to measure the pure overhead of the queue. So, nothing very interesting here. Both concurrent collections perform better than the lock-based one as expected, but there's not a lot to choose between the ConcurrentQueue and my SPSC queue. I was a little disappointed, but then, the .NET Framework team spent a lot longer optimising it than I did. So I dug out a more powerful machine that Red Gate's DBA tools team had been using for testing. It is a 6 core Intel i7 machine with hyperthreading, adding up to 12 logical cores. Now the results get more interesting. As I increased the number of producer-consumer pairs to 6 (to saturate all 12 logical cores), the locking approach was slow, and got even slower, as you'd expect. What I didn't expect to be so clear was the drop-off in performance of the lock-free ConcurrentQueue. I could see the machine only using about 20% of available CPU cycles when it should have been saturated. My interpretation is that as all the cores used atomic memory operations to safely access the queue, they ended up spending most of the time notifying each other about cache lines that need invalidating. The sync-free approach scaled perfectly, despite still working via shared memory, which after all, should still be a bottleneck. I can't quite believe that the results are so clear, so if you can think of any other effects that might cause them, please comment! Obviously, this benchmark isn't realistic because we're only measuring the overhead of the queue. Any real workload, even on a machine with 12 cores, would dwarf the overhead, and there'd be no point worrying about this effect. But would that be true on a machine with 100 cores? Still to be solved. The trouble is, you can't build many concurrent algorithms using only an SPSC queue to communicate. In particular, I can't see a way to build something as general purpose as actors on top of just SPSC queues. Fundamentally, an actor needs to be able to receive messages from multiple other actors, which seems to need an MPSC queue. I've been thinking about ways to build a sync-free MPSC queue out of multiple SPSC queues and some kind of sign-up mechanism. Hopefully I'll have something to tell you about soon, but leave a comment if you have any ideas.

  Read the article

• How to Share Files Between User Accounts on Windows, Linux, or OS X

  - by Chris Hoffman

  Your operating system provides each user account with its own folders when you set up several different user accounts on the same computer. Shared folders allow you to share files between user accounts. This process works similarly on Windows, Linux, and Mac OS X. These are all powerful multi-user operating systems with similar folder and file permission systems. Windows On Windows, the “Public” user’s folders are accessible to all users. You’ll find this folder under C:\Users\Public by default. Files you place in any of these folders will be accessible to other users, so it’s a good way to share music, videos, and other types of files between users on the same computer. Windows even adds these folders to each user’s libraries by default. For example, a user’s Music library contains the user’s music folder under C:\Users\NAME\as well as the public music folder under C:\Users\Public\. This makes it easy for each user to find the shared, public files. It also makes it easy to make a file public — just drag and drop a file from the user-specific folder to the public folder in the library. Libraries are hidden by default on Windows 8.1, so you’ll have to unhide them to do this. These Public folders can also be used to share folders publically on the local network. You’ll find the Public folder sharing option under Advanced sharing settings in the Network and Sharing Control Panel. You could also choose to make any folder shared between users, but this will require messing with folder permissions in Windows. To do this, right-click a folder anywhere in the file system and select Properties. Use the options on the Security tab to change the folder’s permissions and make it accessible to different user accounts. You’ll need administrator access to do this. Linux This is a bit more complicated on Linux, as typical Linux distributions don’t come with a special user folder all users have read-write access to. The Public folder on Ubuntu is for sharing files between computers on a network. You can use Linux’s permissions system to give other user accounts read or read-write access to specific folders. The process below is for Ubuntu 14.04, but it should be identical on any other Linux distribution using GNOME with the Nautilus file manager. It should be similar for other desktop environments, too. Locate the folder you want to make accessible to other users, right-click it, and select Properties. On the Permissions tab, give “Others” the “Create and delete files” permission. Click the Change Permissions for Enclosed Files button and give “Others” the “Read and write” and “Create and Delete Files” permissions. Other users on the same computer will then have read and write access to your folder. They’ll find it under /home/YOURNAME/folder under Computer. To speed things up, they can create a link or bookmark to the folder so they always have easy access to it. Mac OS X Mac OS X creates a special Shared folder that all user accounts have access to. This folder is intended for sharing files between different user accounts. It’s located at /Users/Shared. To access it, open the Finder and click Go > Computer. Navigate to Macintosh HD > Users > Shared. Files you place in this folder can be accessed by any user account on your Mac. These tricks are useful if you’re sharing a computer with other people and you all have your own user accounts — maybe your kids have their own limited accounts. You can share a music library, downloads folder, picture archive, videos, documents, or anything else you like without keeping duplicate copies.

  Read the article

• Copy machine security issues.

  - by David Nudelman

  I am involved on a project to talk to communities about the risks of posting online content is social networks. But this time I was really impressed how far security concerns can go. This video from CBS news talks about security risks related to corporate fax/printers and scanners. It was very clear that when they got the machines they selected the machines by previous owner and they were not random machines, but still, I will never scan from my company machine again. I guess the price of multifunction printers will go up if this video goes viral. Regards, David Nudelman

  Read the article

• Environment naming standards in software development?

  - by Marcus_33

  My project is currently suffering from environment naming issues. Different people have different assumptions as to what environments should be named or what the names designate, and it's causing confusion when discussing them. I've done a bit of research and I haven't found any standards out there. The terms include "Local", "Sand", "Dev", "Test", "User", "QA", "Staging" and "Prod" (plus a few more that different people have asked about) I'm not looking for just opinions, though if there's one out there that "everyone" has I'll take it - I'm trying to find definitions advanced by some sort of authority, even if it's unofficial. Here's the environments we currently use: Environment on the developer's PC Shared Environment where developers directly upload code to self-test Shared Environment where standards and functionality are tested by QA people Shared Environment where completed and QA-checked code is approved by project requesters Environment that mirrors the final environment as a final check and to prepare for deployment Final Environment where code is in use I know what I'd call them, but is there some sort of standard on this? Thanks in advance.

  Read the article

• LAN visibility and network sharing

  - by takeshin

  I have two Ubuntu machines with wifi network cards configured as DHCP interfaces. machine1: inet addr:192.168.168.105 Bcast:192.168.168.255 Mask:255.255.255.0 Gateway: 192.168.168.252 machine2: inet addr:192.168.168.104 Bcast:192.168.168.255 Mask:255.255.255.0 Gateway: 192.168.168.252 They are connected to the router: inet addr: 192.168.168.252 Internet connection from the router is accessible on both of the machines. How to share files between those two? I have already tried few ways (eg. samba), but it looks like the machines are not visible to each other: @machine1$: ping 192.168.168.104 From 192.168.168.105 icmp_seq=1 Destination Host Unreachable What do I need to configure? Apparmor? Firewall?

  Read the article

• Can not resolve hostnames on Xubuntu computers

  - by P Curtis

  I have a network of computers which has been running for many years. I have changed two of those to Xubuntu 11.10 and found I can no longer connect by ssh using the host-name from any other machine. I can connect and ping by IP although ping is very slow in one case (~200ms). All other machines are fine including another with Ubuntu 11.10. Host-name resolution works from Xubuntu machines to other networked machines. I am using wins resolution and have checked settings in /etc/nsswitch.conf are the same as my working Ubuntu systems. What is different in Xubuntu networking that I might have missed?

  Read the article

• Mount exFAT partition in virtual guest machine

  - by Alain Jacomet

  I have a real Ubuntu 12.10 installation being virtualized under a Windows 8 host, by using the VBoxManage.exe internalcommands createrawvmdk method. I'm trying to mount an exFAT partition into the virtualized machine (which is the partition of shared work files), but even though I have fuse-exfat installed, and the partition is perfectly mountable while booting entirely into Ubuntu, I can't mount it while virtualizing it. 1- If I make a full vmdk image of the HDD, including all partitions, Ubuntu 12.10 "sees" the partition, and trying to mount it throws this error: Image: http://i.stack.imgur.com/AyUSn.png 2- If I make a machine with only the linux partitions, + the exFAT partition. Again Ubuntu "sees" the partition and the result is: Error: fsync failed Image: http://i.stack.imgur.com/u4SkC.png 3- If I make a machine with only the linux partitions, and try to mount it, Ubuntu doesn't "see" the partition, and I get this error: Image: i.stack.imgur.com/q1hz5.png I've tried using the VirtualBox' "Shared Folders" functionality but even though I install the "Guest Additions", the system doesn't seem to recognize the shared folder: Image: i.stack.imgur.com/yLU0E.png Help? Thanks!

  Read the article

• Why can't I get a working session with vnc4server

  - by ysap

  We have a couple of (identical) Ubuntu 11.10 machines, configured with gnome-classic, which we use as remote servers, and let our clients log into personal user accounts we create for them using vnc4server. We configured all the machines in the same way, following a short manual we compiled, describing how to download, install and prepare a few tools and our software. The connection usually works fine, but today I set up a fresh machine, and experienced problems. After installing vnc4server, I ran vncpasswd and copied the following startup file to ~/.vnc/xstartup: #!/bin/sh unset SESSION_MANAGER unset DBUS_SESSION_BUS_ADDRESS gnome-session --session=gnome-classic & [ -x /etc/vnc/xstartup ] && exec /etc/vnc/xstartup [ -r $HOME/.Xresources ] && xrdb $HOME/.Xresources xsetroot -solid grey vncconfig -iconic & Then, I started vnc4server and used two viewers (the Ubuntu Remote Desktop Viewer and Windows RealVNC Client) in two other machines, but instead of getting my desktop, I see an empty window with a grey-ish background pattern like this: and the cursor is a bold X. What is wrong with the setup and why don't I get a remote session as expected?

  Read the article

< Previous Page | 79 80 81 82 83 84 85 86 87 88 89 90  | Next Page >