- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic to the Top
- Bookmark
- Subscribe
- Printer Friendly Page
Lab Management with iTest
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content
03-09-2010 03:36 PM
A scenario of interest:
- at least two iTest Clients
- one NE (e.g., a test tool we’ll provide).
- and attempt to control that NE by both iTest Clients
We would like to see a scheme/mechanism that would make both iTest Clients AWARE they are accessing the same NE (blocking, scheduling function, etc… that would be a bonus if available).
Again, the concept is to have multiple iTests available with access to many different NEs… that is already proven to work very well. Now, the question is how to provide all of this flexibility while avoiding folks tripping over each other (e.g., accessing a running test sets already in use by somebody else).
Re: Lab Management with iTest
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content
03-10-2010 09:30 AM
As with any lab management solution, you have to make sure that reservation scheme you set up is used by everyone (automated tests and human beings). anyone not using the reservation system and accessing the equipment directly can cause problems. Trying to set up a perfect reservation system where no one as access to the equipment till they reserve requires huge investment in hardware, firewalls and technology and make them work together. However, in a lightweight version, where you are willing to accept that if someone bypasses the "rules of accessing lab devices", and it is OK, one can set up some very simple reservation systems, For example:
1. You can call a procedure which iterates over list of devices in the testbed and creates lock files on a central share. If lock file exists, it waits for lock file to disappear before proceeding. At the end of execution, it can clear lock files. You can use timestamp of the lock file to determine if certain lock files were left because a testcase got killed and did not return the reservation.
2. You can use database session to add/remove reservation records - you can create your own schema and poll the database to see if someone has already added a reservation record. This logic can be encapsulated in a procedure to be used by all testcases.
There are many other schemes one can come up with for very simple reservation systems.
