Troubleshooting errors
Hi - I'm new to PDQ Deploy but so far love it. I have been using some self-written powershell scripts to do updates til now, so overall I'm familiar with the process. I have a few questions. First of all, is there a decent way to check existing version before upgrading as part of the upgrade task? For example: Java 7 will update other versions of 7, but if a computer has 6, and you put 7 on it won't automatically remove the prior version, so it ends up with two different ones, so my thought is if existing version is x, remove before deploy.
Also, what is the best way to handle errors? I deployed shockwave for example and had some 1603 errors. From prior experience, I have found that what is causing this is a previous version that doesn't remove right, so I generally just delete the old registry keys when I get that, but is there a way to run a certain action if you get a specific error?
I'm sure I'll have more questions, but that is a good place to start.
Comments
Hi John,
To do something like this you would want to add additional steps to your PDQ Package. A good example of this is the ALTERNATE package for Java (available in the Package Library). In the image below you can see that Java 7 Update 25 doesn't actually get installed until Step 27. The previous 26 steps were preparing the target for the install by killing browsers and deleting certain registry keys. **NOTE: This is generally not required for a Java installation. The Alternate package is meant to install Java on targets that have a previously corrupted installation**
There is a package available to uninstall Java 6. You can find this in the Package Library. One thing you could do is run a package which calls the Uninstall package as part of a Nested Step. The image below is from the standard Java package (not the Alternate). I have simply added a new Nested Package step (Step 2) which calls the Uninstall Java 6 package. This happens before Java 7 gets installed.
No, there isn't a native way to run a certain action based on an error. In order to get that granular you'd need to place that type of logic in a script and then call that script instead.
Thank you for your information. I have not tested this, but after seeing a little bit more information, I created an Uinstall task that simply runs the uninstall command line string, hopefully that is fine. As far as your alternate install, I assume there is no issue with running all those uninstall strings prior to running the installer, you just make sure that it doesn't stop the process if that fails?