We have a problem with an Azure website that intermittently fails with this error:
[SqlException (0x80131904): A network-related or instance-specific error occurred while establishing a connection to SQL Server. The server was not found or was not accessible. Verify that the instance name is correct and that SQL Server is configured to allow remote connections. (provider: SQL Network Interfaces, error: 26 - Error Locating Server/Instance Specified)]
and
HttpException (0x80004005): Unable to connect to SQL Server database.
By intermittent I mean that you can start a new browser session and it'll be OK again, then later in the day it'll fail.
There's lots of advice online for this error but it all involves setting up your connectionstring correctly or fixing roleManager or membership in the web.config. None of these solutions would seem to be compatible with intermittent errors on our site (ie. if our connectionstring or web.config were incorrect presumably the site would always fail).
It may be relevant that we had an existing site foo.azurewebsites.net and codebase and we switched to bar.azurewebsites.net and substantially changed the codebase (though starting with the same original files). We've also added some simple Role admin code. Is it possible that because of caching the new site is sometimes trying to connect to the old site's database (now gone)?
But we've had one user laboriously help us out, deleting from his cache anything which was relating to the "old" site ... which fixed his problem ... but next day the problem returned for him.
Update
Recently I was sat here with 2 side by side Chrome browser sessions (different user logins), hitting the site again and again. One session was getting 100% error, the other 0% errors. But I can't reproduce it now. No errors at all for me. But users still saying they're getting a tremendous error rate of 80% to 90% of the time.
Update
It's down again this morning (for one browser session), however many times I try to refresh. A different browser window/identity I fired up alongside it is fine.
Update
Perhaps I have the same problem asked here. Deleting cookies seems to fix it in my case, just as Mark Heath documents. Currently trying the answer Mark posted himself there to see if it helps my situation too.
Presuming you are using Entity Framework
Warning: this works only in EF6+. If you are in EF5 and have this problem, consider updating - it is easy.
If you have intermittent database connection problems in Azure, you should implement retry policy. You can do it via SqlAzureExecutionStrategy. This is described here in detail: http://msdn.microsoft.com/en-us/data/dn456835.aspx
Here is how to enable this:
public class MyConfiguration : DbConfiguration
{
public MyConfiguration()
{
SetExecutionStrategy("System.Data.SqlClient", () => new SqlAzureExecutionStrategy());
}
}
and then you'll need to decorate your DbContext with configuration attribute:
[DbConfigurationType(typeof(MyConfiguration))]
public class MyDbContext : DbContext
{
// blah
}
If you have manually initiated transactions, you'll need to disable retry-policy. For that you'll need to change MyConfiguration to look like this:
public class MyConfiguration : DbConfiguration
{
public MyConfiguration()
{
this.SetExecutionStrategy("System.Data.SqlClient", () => SuspendExecutionStrategy
? (IDbExecutionStrategy)new DefaultExecutionStrategy()
: new SqlAzureExecutionStrategy(1, TimeSpan.FromSeconds(30)));
}
public static bool SuspendExecutionStrategy
{
get
{
return (bool?)CallContext.LogicalGetData("SuspendExecutionStrategy") ?? false;
}
set
{
CallContext.LogicalSetData("SuspendExecutionStrategy", value);
}
}
}
And wrap your transaction calls like this:
MyConfiguration.SuspendExecutionStrategy = true;
// start transaction
// do transaction stuff here
// commit/rollback transaction
MyConfiguration.SuspendExecutionStrategy = false;
Code shamelessly stolen from here: http://msdn.microsoft.com/sv-se/data/dn307226.aspx
Related
Our application is using CefSharp version 73.1.130. The issue only occurs on a small number of internal workstations within our organization. Worth noting, we are also seeing the same error with CefSharp version 92. Another strange thing is that it the issue is consistent, but only when the web apps are launched through certain navigations. Other navigations work consistently for these users.
We use RegisterJsObject to register a javascript object with browser. If I understand correctly, asynchronous binding is preferred moving forward.
The issue presents as strange/unexpected behavior in the hosted web application due to failure to retrieve context from the host WinForms application. The behavior would suggest a failure to register/bind the js object with the RegisterJsObject method. However, that method is not throwing an exception.
Enabled Cef logging showed the following error:
ERROR:JavascriptRootObjectWrapper.cpp(34)] IBrowserProcess is null, unable to bind object
After looking into the code, it appears the location that the value pointed to by "IBrowserProcess" is set is in WcfEnabledSubProcess::OnBrowserCreated (https://github.com/cefsharp/CefSharp/blob/cefsharp/73/CefSharp.BrowserSubprocess.Core/WcfEnabledSubProcess.cpp). I was able to build CefSharp and add additional logging to that method.
On my workstation (I'm not affected by the issue), I get through OnBrowserCreated with no exceptions. However, on my coworkers workstation I see the following line is failing:
...
channelFactory->Open();
auto browserProcess = channelFactory->CreateChannel();
auto clientChannel = ((IClientChannel^)browserProcess);
try
{
clientChannel->Open(); <-- FAILS
browser->ChannelFactory = channelFactory;
browser->BrowserProcess = browserProcess;
}
catch (Exception^)
{
}
}
With the error:
There was an error reading from the pipe: The pipe has been ended. (109, 0x6d)
Has anyone seen this issue before? I'm not sure how much this will help, but does anyone know if it's possible to enable WCF tracing with the CefSharp.BrowserSubProcess.exe. I have been trying this, but no luck so far.
I created a visual studio 2012 MVC4 App. I am testing the "publish" functionality by right clicking the project and choosing publish. I followed the instructions here. I can connect to the remote web server and the folders get published to the correct folder, except the content folder for some reason.
When I run browse to the remote web server it prompts me for login so the app is working. However, the migrations never happened. The only tables created are the simplemembership tables, so I know the web server is connecting to the remote db server. No other tables are created and the seed method doesn't run. I seed the roles and a default user.
I checked the box in publish settings that says "Execute Code First Migrations (runs on application start)"
Everything works fine on my localdb connection string for local testing. Just can't figure out how to create db from existing migrations and seed when I publish to live site, note I will only seed once. Is there a way to specify which migrations to run again? I can copy the migrations and run on the database server but why the extra step?
EDIT:
When adding the database.setinilizer to my context I now get an error saying some of my fields in my userprofile table are not there, I use simple membership. This error occurs on the first page load after web publish, then on proceeding page loads I get an error The "WebSecurity.InitializeDatabaseConnection" method can be called only once.
HOwever, it does create my simplemembership tables now but my migration for all other tables never runs, that is why I am missing the additional user profile fields.
EDIT:
Basically I am not checking if websecurity is initialized prior to calling WebSecurity.InitializeDatabaseConnection so that resolved that issue. Now I have it partially working. The app creates the simplemembership tables fine but since I added tables to the UserProfile table I can't seed until I change them. So instead I manually create the userprofile table and have the app create the rest of the tables. Then I comment out the userprofile table in my initial migration. After this when I sign in it will then create the rest of my tables.
Open issue is how to get my database migration to run prior to the simplemembership initialization?
To get migration work on remote server, you need to add use SetInitializer in you Context class first :
static MyDatabaseContext()
{
Database.SetInitializer(new MigrateDatabaseToLatestVersion<MyProjectContext, Migrations.Configuration>());
}
And in you Migration Configuration you need to add this code :
public Configuration()
{
AutomaticMigrationsEnabled = false;
AutomaticMigrationDataLossAllowed = false;
}
I don't select the "Execute Code First Migrations (runs on application start)", and just after setting initialization in MyProjectContext, it does the migration for me.
If you have done by here, for seed your data, you can do same as below in your Migration configuration class:
protected override void Seed(MyProject.Models.MyProjectContextcontext)
{
DoSeed(context);
}
private void DoSeed(MyProjectContext context)
{
var users = new List<User>
{
new Site { UserId = 1, Name = "TestUser", IsDeleted = false },
new Site { UserId = 2, Name = "TestUser2", IsDeleted = false }
};
users.ForEach(s => context.Users.AddOrUpdate(s));
context.SaveChanges();
}
I have not selected the "Execute Code First Migrations (runs on application start)" on deploy Profile.
For some more details on this see this:
This link
and
This link
Hope this helps(as it worked for me), otherwise please add any error if there is, when you deploy you app.
I think the issue is ,Because of the fact that as long as you have not tried to access data
or create any data from/in site, which needs to connect to database, the migration and seeding
will not run"
And the reason for running migration on your site after logging into the site,
would be because your site need to be authorised in all pages, or the page that you want
to see data in.
If you have a page example home page, that does not authorization to access to the page,
you will see the page and if in this page there is some data that needs to be fetched from
data base, you may see the migration runs.
I found this from this Deploy it to IIS, but not sure if it is the reason.
Please let me know if your migration still has not ran if you browse your home page that
has data and no authentication needed for this data access.
I have a site built using MVC4 which is getting content out of a database and it all works fine.
If I then rebuild the solution and try to refresh the page to check my changes, I will always get the SQL Network Interfaces, error: 26 saying that I cannot connect to the server.
However, if I then browse to my homepage and then back to the page I was looking at it will work fine.
Does anyone know what could cause this problem as it is really annoying
EDIT
Further to this I have found it is when the AuthorizationContext filterContext is being loaded after the rebuild that it cannot connect to the db
EDIT 2
As with neil below I have found that I only get the problem if I try to access a page that has had a role assigned to it
I'm seeing the exact same problem and can trace it to the .ASPXAUTH session cookie in the browser. Delete that cookie and the database error goes away until the next rebuild.
The error occurs regularly if you are authenticated and then rebuild the project and try to browse any page that either:
Requires authentication
Makes a call to the User object (e.g. #if (User.IsInRole("Administrators")))
If you have the AuthorizeAttribute filter set in App_Start/FilterConfig.cs you'll get this on every page.
This seems to be new behavior following the most recent Patch Tuesday updates. Previously, I was seeing weird behavior where I would remain logged in but I would loose my roll membership. After the most recent patches, it seems Simple Membership chokes when it gets a bad .ASPXAUTH cookie (invalid because of the rebuild).
I've got the correct connection string in InitializeSimpleMembershipAttribute.cs but it's like Simple Membership is defaulting to something else in this one instance.
Note that I've moved the Simple Membership databases from the original (localDb) to a full-fledged (local) SQL Server instance. Don't know why that would matter, and it works fine in all other cases.
Update:
I've also tried making the connection string name the same as the EF context name (e.g. "ProjectContext") on the theory that it is defaulting to the standard convention, but that made no difference. I am explicitly identifying the connection string name in all my context class constructors (using the : base("connectionString") syntax) and Simple Membership is able to find the right connection string all other times.
Update 2:
The problem only occurs after rebuild when accessing a page protected by role. A simple [Authorize] won't trigger it. You need something like [Authorize(Role="Admin")]. I've replicated this on a new MVC 4 project with no other modifications, using the default (localDb) database. Clear the cookie and that same user can access the protected content w/o any problems. I believe this is a core .NET or MVC bug and needs to be reported as such.
This happened to me while rebuilding the application when I was logged in the browser.
Deleting cookies fixed the problem for me.
When using SimpleMembership, this bug occurs with any use of Roles-not just in the controller, but also when I use:
if(Roles.IsUserInRole(rolename){...
I am late to the game with this answer, but I added [InitializeSimpleMembership] to my home controller and I think that fixed it.
I have BasePage.cs class which is being used by other .cs files instead of System.Web.UI.Page (public partial class page : BasePage).
I am using it for opening and closing SQL connections to make sure every SQL connection gets closed.
Code looks like this
{
public class BasePage: System.Web.UI.Page
{
public SqlConnection globalConnection;
protected override void OnInit(EventArgs e)
{
globalConnection = new SqlConnection();
globalConnection.ConnectionString = ConfigurationManager.ConnectionStrings["kontemiConnectionString"].ToString();
globalConnection.Open();
}
protected override void OnUnload(EventArgs e)
{
if (globalConnection != null)
{
globalConnection.Close();
}
}
}
}
So far it has worked really well for me. It means that every time a connection opens it also gets closed. Or at least I thought so.
My question is whether this solutions is bulletproof and every single connection gets closed in case there is some processing error during code execution. When tracing this code, if I on purpose create error 500 it always goes to OnUnload event and gets closed.
So, do you think is this execution safe?
(To stop discussion whether I shouldn't open SQL when I actually need it, answer is that every page which uses BasePage also opens a SQL connection.)
Opening and closing SQL connections where you are using them is better practice. Under the hood, the CLR manages connection pooling anyway - reusing connections and closing them when it sees fit. Doing lots of open and closes on connections with the same connection string doesn't add up to as much overhead as you might expect.
I might be wrong, but my guess would be that the connection does not get closed if the page lifecycle is ended through an exception before OnUnload fires. The very least you could do to prevent this is make sure you catch all exceptions on a global 'last-chance' level, and close the connection there. I still think using connections locally, ideally with using blocks, is the better solution, because it doesn't keep connections open longer than it needs to, and you don't have to worry about closing them (the semantics of using will do the work for you).
I work at a college and have been developing an ASP.NET site with many, many reports about students, attendance stats... The basis for the data is an MSSQL server DB which is the back end to our student management system. This has a regular maintenance period on Thursday mornings for an unknown length of time (dependent on what has to be done).
Most of the staff are aware of this but the less regular users seem to be forever ringing me up. What is the easiest way to disable the site during maintenance obviously I can just try a DB query to test if it is up but am unsure of the best way to for instance redirect all users to a "The website is down for maintenance" message, bearing in mind they could have started a session prior to the website going down.
Hopefully, something can be implemented globally rather than per page.
Drop an html file called "app_offline.htm" into the root of your virtual directory. Simple as that.
Scott Guthrie on the subject and friendly errors.
I would suggest doing it in Application_PreRequestHandlerExecute instead of after an error occurs. Generally, it'd be best not to enter normal processing if you know your database isn't available. I typically use something like below
void Application_PreRequestHandlerExecute(Object sender, EventArgs e)
{
string sPage = Request.ServerVariables["SCRIPT_NAME"];
if (!sPage.EndsWith("Maintenance.aspx", StringComparison.OrdinalIgnoreCase))
{
//test the database connection
//if it fails then redirect the user to Maintenance.aspx
string connStr = ConfigurationManager.ConnectionString["ConnectionString"].ConnectionString;
SqlConnection conn = new SqlConnection(connStr);
try
{
conn.Open();
}
catch(Exception ex)
{
Session["DBException"] = ex;
Response.Redirect("Maintenance.aspx");
}
finally
{
conn.Close();
}
}
}
You could display a message to people who have logged in saying "the site will be down for maintenance in xxx minutes" then run a service to log everyone out after xxx minutes. Then set a flag somewhere that every page can access, and at the top of every page(or just the template page) you test if that flag is set, if it is, send a redirect header to a site is down for maintenance page.
What happens now when the site is down and someone tries to hit it? Does ADO.NET throw a specific exception you could catch and then redirect to the "website down" page?
You could add a "Global.asax" file to the project, and in its code-behind add an "Application_Error" event handler. It would fire whenever an exception is thrown and goes uncaught, from anywhere in your web app. For example, in C#:
protected void Application_Error(object sender, EventArgs e)
{
Exception e = Server.GetLastError().GetBaseException();
if(e is SqlException)
{
Server.ClearError();
Server.Transfer("~/offline.aspx");
}
}
You could also check the Number property on the exception, though I'm not sure which number(s) would indicate it was unable to connect to the database server. You could test this while it's down, find the SQL error number and look it up online to see if it's specifically what you really want to be checking for.
EDIT: I see what you're saying, petebob.
The "offline.html" page won't work if the user was already navigating within the site, or if he's accessing the site from a bookmark/external link to a specific page.
The solution I use is to create a second web site with the same address (IP or host header(s)), but have it disabled by default. When the website is down, a script deactivates the "real" web site and enables the "maintenance" website instead. When it comes back online, another script switches back to the "real" web site.
The "maintenance" web site is located in a different root directory, with a single page with the message (and any required images/css files)
To have the same message shown on any page, the "maintenance" web site is set up with a 404 error handler that will redirect any request to the same "website is down for maintenance" page.
A slightly more elegant version of the DB check on every page would be to do the check in the Global.asax file or to create a master page that all the other pages inherit from.
The suggestion of having an online site and an offline site is really good, but only really applicable if you have a limited number of sites to manage on the server.
EDIT: Damn, the other answers with these suggestions came up after I loaded the page. I need to remember to refresh before replying :)
James code forgets to close the connection, should probably be:
try
{
conn.Open();
}
catch(Exception ex)
{
Session["DBException"] = ex;
Response.Redirect("Maintenance.aspx");
}
finally
{
conn.Close();
}
Thanks for the replies so far I should point out I'm not the one that does the maintenance nor does I have access all the time to IIS. Also, I prefer options where I do nothing as like all programmers I am a bit lazy.
I know one way is to check a flag on every page but I'm hoping to avoid it. Could I not do something with the global.asax page, in fact, I think posting has engaged my brain:
Think I could put in Application_BeginRequest a bit of code to check the SQL state then redirect:
HttpContext context = HttpContext.Current;
if (!isOnline())
{
context.Response.ClearContent();
context.Response.Write("<script language='javascript'>" +
"top.location='" + Request.ApplicationPath + "/public/Offline.aspx';</scr" + "ipt>");
}
Or something like that may not be perfect not tested yet as I'm not at work. Comments appreciated.