I am noticing that while still largely successful, most of the trigger-based policies in my environment produce failed or otherwise unsuccessful push attempts on client machines within the specified scope.
Take for example the case of a Dropbox update that I created this morning. This policy is scoped to an existing Smart Computer Group for those requiring the most recent release of the application. The total scope of machines started an hour ago was targeted at 428 clients. As of right now, 115 clients have successfully installed the update while 16 have failed. All 16 clients are failing with the following error code.
Executing Policy Dropbox 3.14.7 Push... Downloading Dropbox 3.14.7.pkg... Downloading <JSSPackageLink>... The network connection was interrupted while downloading the package from <JSSPackageLink>. Attempting to reconnect... Downloading Dropbox 3.14.7.pkg... Downloading <JSSPackageLink>... Error: Dropbox 3.14.7.pkg is not available on the HTTP server. Running Recon... Retrieving inventory preferences from <JSSLink>... Locating accounts... Locating package receipts... Searching path: /Applications Locating software updates... Locating printers... Gathering application usage information...
I used to believe that the "the network connection was interrupted while downloading the package" error suggested that the client machine had lost network connectivity during the download but given the frequency that this was occurring, I decided to investigate. What I found was that I was able to ping and perform an SSH session onto the machine with the provided IP address in our JSS which suggests to me that the client never lost connectivity. Has anyone experienced this or something similar in their JSS environment? If so, could you please weigh in on this?