I have a couple of continuous web jobs which are currently stopping. I know the reason for this: I haven't got "Always On" turned on in my web site settings and the sites eventually unload.
The problem is that I also use New Relic for site monitoring, and their agent that collects data stops working when "Always On" is switched on.
So a catch-22: have the webjobs work but no performance and error monitoring, or lose the webjobs and have the monitoring.
Trying to keep HTTP traffic going to my site by loading a URL doesn't seem to stop the webjob unloading.
Does anybody know how I could get the best of both worlds?
This ended up being simpler than I thought once I stopped thinking about how I could get both things to work on the one web site when there's actually no need!
In the end it was far easier just to create a new, empty web site to house the webjobs. These could be kept "Always On" without affecting the main web site.
This is a catch 22. The main problem is "Always On" has to be turned off because of the way that Azure Web Sites work. There is nothing that can be configured and/or coded into the agent that will affect the way it works at this time. Microsoft is performing some updates to Azure in the near future, and we are hopeful that some of the changes will help with issues like this.
About the only thing that I can suggest is to create some automation to not only load a URL, but perform whatever action on the site is necessary to keep the webjob active. It is possible that the New Relic Synthetics option would be able to help with this when it is released, but there's no way to say for sure at this time. In the mean-time you might want to look at other automation tools, or possibly building a Ruby application, that can open a page and perform the necessary actions to keep the application active.
Sorry there's not better news for you.
Related
We have moved our website which uses typo3 from on-prem to Azure cloud. We setup a Front door with firewall protection which is different from the previous setup.
Since day one when I log in I can do some stuff for a short while (like 4-5 minutes) and then it kicks me out to login screen.
Another example is when I'm logged in, I open a new tab and check some other sites then go back to typo3, again I'm logged out. Need to log in again.
I lost some of my posts while adding some additional info from other websites.
Any ideas?
I had a similar issue. I resolved it with replacing lockIP in the install tool from 4 to 0.
Note, this is a temp solution so you can keep working, but you really need to find out why this is happening.
Best regards
Every time I try to connect to any SharePoint site running on my customer server I get the following error.
An error occurred accessing your Microsoft SharePoint Foundation site
files. Authors - if authoring against a Web server, please contact the
Webmaster for this server's Web site. WebMasters - please see the
server's application event log for more details.
I already checked the sharepoint options in CA and in every site in the mentioned server. Everything seens correct but the error persists.
It looks like there is no direct relation to SharePoint and something else in the server but I was not able to determine what could cause such behaviour.
Hope someone here has something to say.
Well... After a very long and painful headache we were able to find out what was causing this malfunction.
It was a .net agent from New Relic. This agent is used to analyse traffic, page load time and some other cool things. It basicly add some javascript to the head session of every response our IIS make and this code send some data to New Relic servers that will be processed to build some reports about the applications running in the IIS.
In the end, I just disabled it and SPD turned back to life.
Thanks.
PS: Boland was in the right path. I were able to find out the solution using Fiddler to analyse the responses from IIS. Thanks.
You have to check the event log. Do you see any error there? If not, it must be a network issue. Firewall is rare, because SPD works on TCP/80.
I am launching a new redesigned website on windows hosting. I am wondering what is the best way to launch this new website without having any downtime on the existing one?
MY only fear is having a user go to visit a page and it's not there or the supporting files are not uploaded yet.
One of the simplest ways to handle this is to put a load balancer or proxy server in front of the application server. Then set up another application server with the new code. Once it is ready, you can change the proxy server to point to the new application server with the new code. Once you are sure nobody is using the old application server, you can shut it down. This, of course, relies on your ability to get that setup in place. If you are on a budget, you might be able to do it all on a single box. For instance, you could use nginx as a reverse proxy to your application on the same box. Getting that in place could potentially cause a tiny window of downtime - not sure if that's acceptable. Then you might be able to set up the new application on the same box with a different port - again, I'm not sure if that would work for your setup. Anyway, the reverse proxy approach is a pretty common one, and one of the great reasons for deploying to the cloud. You only pay for the short period of time when you need both boxes.
You should make sure that your new website launches all at once and that you set up the proper redirection rules for all previous pages. Once you are launching the new website, pick a time at night where you have low traffic volume, and simply upload all the new code at once to the webserver. This eliminates the fear you have of the "supporting files not uploaded yet". One of the key things to do is make sure all your old pages redirect and map over to new pages on the site just in case anyone clicks into your site using external links.
Two good resources to read:
http://www.rise.net/blog/ideal-way-launch-website-rebrand
http://googlewebmastercentral.blogspot.com/2008/04/best-practices-when-moving-your-site.html
The best methos is to upload the site via FTP, and if you have RD access to login into the Windows server and to copy the new site for a few seconds. In this way you will not have any downtime as when you directly upload the site via FTP.
I have had my blog running on Orchard in Azure for, I dunno, a few months probably. All has been well. I have about 10 content items. It is a small site. I'm running it in Shared mode. Browsing the site is fine, it is fast and all is great. But today I have been trying to edit some posts and add some new ones. My per minute CPU usage is just going crazy and keeps crashing the site and hence I cant save anything. Pressing publish just destroys the site.
I'm not upgrading to reserved mode like it keeps recommending for a fucking little blog with about 3 viewers.
Any ideas why the cpu usage could be going so crazy?
Error logs are pretty much empty, occasional error from disqus but that is only when I am loading blog posts...
UPDATE 1:
Removed disqus just in case. Still failing miserably with massive loads to publish content items.
UPDATE 2: Kinda strange... error logs say A tenant could not be started: Default. Sequence contains more than one element. I think it is talking about routes.
at Orchard.Mvc.Routes.StandardExtensionRouteProvider.d__a.MoveNext() in c:\Users\sebros\My Projects\Orchard\src\Orchard\Mvc\Routes\StandardExtensionRouteProvider.cs:line 24
You should check that extensions monitoring is disabled. It creates lots of FileSystemWatcher instances in order to get dynamic compilation responsive to live files modifications but it's unnecessary in production environments.
Look at "Disabling the Dynamic Module Loader" section on this page: http://docs.orchardproject.net/Documentation/Orchard-module-loader-and-dynamic-compilation
I have recently installed a website project on Azure and that was relatively easy to do thanks to great docs online. We are having a problem with the back office (admin) login though, it's a bit strange, as it didn't happen straight away, I shall elaborate....
When I go to http://www.keelycattschoolofdance.co.uk/umbraco/ it takes me to the login page, which seems fine.
(The url at this point becomes www.keelycattschoolofdance.co.uk/.../login.aspx)
I then login, which works and it starts to display the admin panels but then bounces back to the login and the URL becomes www.keelycattschoolofdance.co.uk/.../login.aspx
I am running this on Azure using the Accelerator, latest version of Umbraco and using SQL CE 4.0 rather than SQL Server.
has anyone seen this before?
Please let me know if you need more information, I shall respond :-)
Thanks in advance
Nick
It sounds like it may be a Session error. This is a problem for Azure and Umbraco because Azure uses round robin load balancing.
The normal way of setting up an Umbraco site using the accelerator would be to have a separate session DB where the shared session state is stored.
Make sure you are using this accelerator for Umbraco on Azure as it's the latest and best.
http://waacceleratorumbraco.codeplex.com/
We have several large production sites using this and I can confirm they all work fine (including the backend editing environment).
Things to watch out for are that you need to follow the instructions closely as they are easy to get wrong :-)
Make sure you don't have dashes in your db names as this makes things go wrong as well.
Finally if you are using more than one web role the Umbraco preview can sometimes yellow screen, as the preview XML may not be on the server that gets to the preview request. This is again because of the round robin load balancing.