I have task to deploy my grails application on remote office. It was far from my current location and it would be waste of time to travelling there to send update.
But I got another problem, we have slow, unreliable and limited quota for internet connection (3rd world country problem). 1 war file is about 55-60 MB.
I also don't want to send my code to them and let them compile and run it.
Could we use SVN/GIT just to send the update of compiled class to them ?
if yes, how to do that ?
If your application is having <Context reloadable="false" /> then
You can simply replace the class files(which are changed) in the location
apache-tomcat/webapps/applicationName/WEB-INF/classes
You could use something like rsync (or robocopy if you're on Windows) to copy to your remote office. That way, for each update, only the changed parts need to be transferred over the network and you should have less data sent. You could rsync the WAR file itself, or a directory which is the WAR file exploded (uncompressed). One might be more efficient with network traffic than the other, but you'd have to test that yourself if you wanted to go that far.
Also, don't forget that Groovy is compiled down to Java class files. So even if they don't have your source code they can decompile the class files and get source out of it (even though it won't be as pretty). This is worth remembering if you have any "secrets" (database password, etc.) in your source code.
I have a server / client project, both written in dart. Now my server starts on port 1337 and when I run my client with the Run in dartium, my static files are served on port 3030 which allows me to debug my client code in the Dart editor.
The problem is that this causes CORS when using AJAX calls. I have properly setup my server to accept other origins (with Access-Control-Allow-Origin) but, for example, cookies aren't sent along.
Now I'm wondering: is there a way to serve my files with my server (running on 1337) and still have the possibility to debug the client side code in the dart editor?
My understanding is that you can debug, but the real problem is that you don't get the expected data back from the server due to missing cookies.
Standard CORS requests do not send or set any cookies by default.
In order to include cookies as a part of the request, besides setting up the server, you need to specify withCredentials property, e.g.:
HttpRequest.getString(url, withCredentials:true)...
You will also need to setup server to provide Access-Control-Allow-Credentials header.
EDIT: it seems that additional issue is that you don't want to have 2 servers, each serving different part of app.
In that case, you can configure DartEditor to launch the URL, instead of files. Go to Run > Manage Launches and add create a new Dartium or Dart2JS launch with specified URL and source directory.
Another option is to select Run > Remote Connection and attach to a running instance of browser or Dart VM.
Caveat: I haven't tried these options, so I can't tell how stable they are.
I'm using Microsoft's Web Deploy Remote Agent service to allow me to easily publish code to the server from within Visual Studio.
The web site I am deploying is using log4net to log messages to log files, and every time I try to deploy a new version of the code, I get this error in Visual Studio stating that the current log4net log file is in use:
An error occurred when the request was processed on the remote
computer. The file 'Web.log' is in use.
The process cannot access 'C:\inetpub\wwwroot\Logs\Web.log' because it
is being used by another process.
I can solve this by going onto the server and doing an iisreset before publishing... but that is kind of defeating the point of 'easy' publishing from Visual Studio :)
Is there some way I can get the publish task to issue an iisreset automatically, or some other way I can work round this?
I kept poking around and found some tidbits around the file being locked in a few other forums. Have you tried adding
<lockingModel type="log4net.Appender.FileAppender+MinimalLock" />
To your <appender> element in the web.config file? From the Apache docs
Opens the file once for each AcquireLock/ReleaseLock cycle,
thus holding the lock for the minimal amount of time. This method of
locking is considerably slower than FileAppender.ExclusiveLock but
allows other processes to move/delete the log file whilst logging
continues.
As far as the performance considerations, I suppose you would need to test if this will affect you or not as I am assuming it really depends on how often you are writing to the log file as to how much this will impact performance. I can't believe that getting/releasing a lock could take all that much time though.
There is a MSDEPLOY provider called recycleApp which is used exactly for this. You can include this in your deployment manifest.
Another option is to use ignoreOnErrors flag which will skip the file in use and continue with the deployment.
I need to update my application from a central server.
The application checks always if it is a correct version, against the server installation.
So when it is not, I need it to update itself.
So how can I copy the EXE if it is running? What solution do I have?
I rename the current running exe to MyTempExe.exe, copy the new exe to the correct location (request elevated privileges if necessary) and then run a separate app to restart the main app. On start up I check for MyTempExe.exe delete it if it's there.
The reason I use a separate app for the restart is I don't have a set time frame for the app to close down and need to wait for it to finish whatever it's doing, on shutdown it writes information to disk about its current state that the updated app will use to resume where the old one left off.
I don't know if it's the best solution but it's the one I use.
As you can see by all the answers there is no set way to do this, so I thought I would add the way we have successfully done this.
We never run an application directly from the network.
We run the application from the local machine and have it copy from the network on startup.
We do this using an application launcher. It downloads an XML file that contains CRC and Version Resource Values for the application files. The XML File is created during the deployment process, in a FinalBuilder Script.
The application then compares the XML File to local content, and copies down needed files. Finally we then launch the application in question. This has worked well for deploying an application that serves around 300 local users. Recently we switch from a file copy to an HTTP download as we found problems with remote user disconnecting drives.
We still still build installations (With Innosetup) to get the basic required files deployed.
Package your app with an installer such as Inno. Download and execute the installer. Have the installer search for and kill your app, or instruct the user to close it. The setup will replace your .exe, and if the app can't be killed or the user is non-cooperative, it'll issue a re-start notice.
Download new EXE to TEMP
Create Batch from EXE, content:
taskkill /PID %process id of running EXE%
copy %new EXE% %running EXE%
%EXE%
all values in %...% are placeholders
execute batch from the running EXE
delete batch
I use TMS TWebUpdate myself, for software updates. The advantage is that there a bunch of extra actions you can put into the script, if you need anything other than plain EXE updates.
I have two components at work the application executable itself and a web-service (SOAP) which provides version details and file downloads.
The application calls a method on the SOAP service to ask for the number of files in the project (project is identified by using the application.exename usually).
The soap service gets its info from an INI file, which has entries like:
[ProjectName]
NumberOfFiles=2
File1=myapp.exe;1.0.0.1
File2=mydll.dll;1.0.0.2
You just update this file at the same time as uploading your new files.
The process of updating the application this:
Get number of files available on the web service
For each file, the application asks for the name and version number from the SOAP server.
The application compares this information to its own version info and decides if the file needs updating, building a local list of files that need updating.
For each file that needs updating the application downloads the file to filename.ext.new
Finally, the application renames all filename.ext to filename.ext.old and renames filename.ext.new to filename.ext and then restarts itself. (No real need for an external app to restart your own program).
Note 1, that you may have to ask for elevation to replace files, depending on where you install your files.
Note 2: be kind to your users, think carefully before you force updates on users.
Note 3: You cannot delete a running exe, but you can rename it and then restart the new version.
Edit===
For some reference data files which cannot contain version information resources, you can have entires like File99=MyDataFile;1.1.2011 the 3 elements to the version number indicates to the client that it should check against the file date/stamp.
You could have a separate update executable whose task is to check the server version, download an updated executable if necessary, and then run the local executable.
Or you could have one executable running in two different modes: 1. on startup, check for an update, if there is one, download the executable to a download directory, run it and quit.
2. The new executable would check if it's running from the installation directory, if not, it would copy itself there, overwriting the old version, start the copy from there, and quit.
My way is the other way round: If a new version is online, promt the user to update. If he want's to (or is forced to...) I end the app and start a new exe (updater). this updater loads the update and replaces the old exe (not running). then it starts the new exe. ready. (You can of course replace other files too.) BUT: Using an Installer like InnoSetup gives you more possibilities and doesn't mix up with the regular uninstaller, so it is really better...
You can do this without running another application. Push the updates to the client from the server while running, storing in a temporary directory on the client. When you want to upgrade move all your running files to another temporary directory, move the new files into the original application directory, and just restart the application using the standard executable name on shutdown.
I upgrade client applications running on unattended machines automatically this way.
I have been able to build rabbitmq server on ubuntu linux. It came already prepackaged and on making, it is able to start as a service. When i got the client source, i failed to make because it appeared like it needed a folder called ./deps/rabbitmq-server. Analysing the code, i find that the author of the client was accessing the same header files as are found in the server, using include_lib("path to rabbit.hrl e.t.c") in his header file called "amqp_client.hrl". I then decided to add rabbitmq_server in the lib dir of erlang so as its paths are automatically added on start up of the vm. But still this didnot help. There is also another folder which the client references called "rabbit_common" for an include folder he assumes would contain all the .hrl files there. Please assist me in building both the client and server on my ubuntu server, for testing.
Also, if anyone has used RabbitMQ server for IMs, please provide some benchmarks and/or your findings on its throughput, speed and number of users. How can it be compared to ejabberd?. How can one create AJAX/Jquery/Javascript clients for Web functionality?
thanks
I hope you had made some progress as far as RabbitMQ and ejabberd are concerned.
Below is a link to an interesting discussion that might be of help.
http://old.nabble.com/AMPQ-vs-XMPP-and-RabbitMQ-vs-ejabberd-td17587109.html