Just wanted to know if anyone has a workaround to this issue. I am trying to patch the Dynatrace server that is running on Redhat. However, in the patching assistant, when I am trying to upload the patch (300mb), due to network instability, it gets disconnected very quickly and the patch fails.
Going to a nearby physical location to the AppMon server is not an option.
Is there a way to SFTP or SSH to put the patch into the Linux box before applying the patch?
There is a backdoor way to add a patch to a dTServer, but it requires several manual steps. And be aware that this is not supported. The vague idea of the steps is to upload and unpack the .dtf file into the same directory that the GUI would use for the upload step. Then edit an .xml file to reflect that the patch was uploaded but not installed. Then run the client and go into the Updates tab and it should show you this patch as ready to be installed but "Inactive'. From this point simply complete the fixpack installation procedure via the dT Client GUI.
I'm concerned that if you're having network problems doing this type of upload, then you would have general client problems and there's something wrong with your network.
But here's the vague steps of how to do this manually:
Note: I have not tested this procedure so full disclaimers come with these steps. Let me reiterate that this is not supported and should something catestrophic go wrong, you could be required to reinstall your dTServer instance
Let us know if it works.
Thanks for the reply! =D
I will need to discuss with the customer because it is a production environment and therefore we need to be cautious.
Just another question, how do you unpack a .dtf file through linux command?
Note: this server hasn't been patched before and is running 6.3.0 version, which is why we are facing problems now.
Sorry, forgot to mention the .dtf unpacking. A .dtf file is really just a zip file. So rename the file to .zip and you can use any standard zip utility. On linux "unzip <filename>.zip".
Given you've never done a update on that system, I suggest you do this procedure on another instance and see the file patterns, and ensure you see the syntax in the .xml file. That .zip file may not exist until you perform the first update operation. Practicing the whole procedure on another test system would be highly recommended given this isn't supported nor tested.
I followed your steps you mentioned for applying patch manually without client..For me it didn't go as planned. So what i did is, i copied the already applied fixpack in 1 of our dev server and moved it into Test server.
I rolled back the current active fixpack in test evnvironment, and it autmatically took the new fixpack after restart.