I have an RPi2 iot-core app which has run fine for months on the previous iot-core os release. I access it regularly via the iot-hub and it writes to Azure blobs ... until the automatic update last week to version 10.0.14393.67 at which time my UWP-headless-app, although still running (I see this in my reference version via the local device portal app), no longer communicates with iot-hub or azure.
Any ideas what going on and what can I do without having to start from scratch?
Related
I am developing an invoice app with Angular + NodeJs + MySql.
The thing is, the app is planned to be used by one employee in his office. No need for online servers.
It is not problematic to deploy the app online, but the internet is unstable in the zone (Latinamerican problem. You may lose connection for hours, and even voltage variations that may shut down the PC).
So the app must be self sufficient to always work offline.
So my questions are:
Can I simply deploy the app offline? Like in local. If that is the case, I would need for everything to be initialized automatically when the user opens the app (server open, database connected...).
If I have no way but to deploy the app online, should I use Firebase? Also, what happen if the internet service shut downs for hours? Is there a way for the database to be available offline and sync when the internet gets back?
You could build the app as an Electron App, then its becomes a locally run program. https://www.electronjs.org/
You can host it anywhere, but turn the app in to a PWA, which means it will work locally in the browser after a successful visit (gets installed with a service worker in browser) For the database it self, you can store data in the browser but some are limited to 5mb of data in the localstorage / sessionStorage / indexdb. Firebase does have some locally cached data. But if the browser is closed it can be lost.
If it needs to run locally i would go the electron route. Its slightly harder to do but it fills out your usecase better.
You can use both ways if you want to be sync like situation you have to hold data if your internet is not working in local storage or indexed db.
and it is fine you can deploy locally also or make one dedicated server which is always on.so any body in same network can use that angular app easily.
Just take care of backup plan when you system corrupt you should have proper backup of database for such scenario.
Created several app packages for a windows store app some time ago in order to show a client the progression of the application. Now I would like to load a much older version of the application onto the device, however PowerShell says that the certificate has expired and thus the application does not install. Is there a way to regen the certificate without recreating the entire app package and/or get around an expired certificate and just manually install the app on the device?
I know what is metro app (just did some research) my question is where metro app runs. By my understanding :
Detktop app: running on your local machine with .net framework, and present data to user, you can launch/close it.
Web app: Publishing on IIS server, running from server render html to client, user can send request and get result.
Now, Metro app: is it just distributed on windows store, user can download/install it and run from local as same as desktop? or run from store, render something to user like web? or maybe some new working styles, can user close it?
Metro Style App was a working name before Microsoft settled on Windows Store Apps.
Windows Store Apps just like desktop apps run on local machine, but are often architected in more modern ways where the app would often connect to web services or push notifications that run in a data center. The UI though usually runs on local machine even if it is written in html/js with winjs.
The biggest difference is distribution - web apps run on the server so you only need to navigate to them in your browser to run the latest version. Desktop apps need to be installed - either with xcopy-style install, Windows installer or perhaps the ClickOnce technology. Store apps are distributed through the store which makes them easier to install and update than desktop apps as well as monetize, but are sandboxed (have limited access to local machine resources) for security.
The Windows Store App Cert kit fails to start.
Normally the flow is:
Start App Cert Kit
Choose Windows Store Application
Choose App
Choose Certs to Run
Run Certs
Save results
Currently, it fails just before 'Choose App'.
I get this result:
I have tried uninstalling and reinstalling, installed all updates, running appcert.exe reset. All to no avail.
I think it may have to do with the same issue I get with the Windows Store app. I have 38 languages installed (our app supports all of them) and the Windows Store app doesn't like it. I've had to escalate with customer support repeatedly for this, and have not had it resolved properly. I wonder if the same issue applies here.
So I was unable to start the UI, but I was still able to run the appcert by manually calling it on the command line as described here. It worked fine. One thing to note is that between each attempt, make sure to run appcert.exe reset.
I have integrated my selenium Webdriver scripts (using TestNG) with Hudson. i invoked my job through ant . My problem is my scripts are not running successfully also the IE browser is not getting opened However Build is creating successfully.
Note -
1) I am triggering build on Hudson from different machine on the same network with administrative access.
2) I have used excel 2007 for developing the script ( data driven framework) however on Hudson server its open office.could that be the issue?
At this moment i am not using selenium grid .please provide any suggestion
The reason for this is because Hudson/Jenkins is running as a Windows service. Recently, Microsoft changed services so that they run on their own invisible desktop. This didn't used to be the case until a few years ago. For this reason, even if you check "interact with desktop" , the desktop that is being referred to is invisble .
So, what you have to do is run Jenkins (not Hudson) as a service only for the master server. For the Selenium tests, you need to run another slave Jenkins server as a foreground shell process and the selenium tests need to be launched from that Jenkins instance instead.
If you think the instructions for setting up a slave are too hard, then you should know that it isn't required that you run as an official configured slave. You can run as a separate slave (make sure port number is different).
Also, if you are running Grids and Nodes on Windows, you might like these scripts I made.