We are trying to deploy Adobe Reader 9.1.1 with 3 language packs.
It is a batch file that calls each of the MSIs individually if the previous install failed. The deployment is 12 files for a total of 145 MB. SWD task is a policy and it points to the HTTP shares (not UNC) as we also have a mangement gateway and that forces us to use HTTP shares when doing SWDs.
We have 1 core server (8.8 SP3), 2 preferred servers, 3800 clients, ~60 different sites scattered all over the world.
We have about 122 computers that are reporting: "Failed to download all additional files for a package". About 1500 computers have received and installed the deployment successfully. The rest are still waiting to get it.
I took a look at one of the ones that failed with this error and it says that the hash did not match for one of the MSIs that were downloaded. I checked the MD5 checksum for this particular MSI on the core and the 2 preferred servers and they all match (EMvWwDJtjVGsFVeHq1BQhg==). However, on this particular PC, the checksum for this MSI was dQrtqRpkdCTXh4ODyM+FzQ== and the MSI instead of being ~20 MB was ~40 MB. This PC was in one of the offices so we can rule out the gateway or vpn issues being a problem.
1) How can I find from what peer this computer downloaded this file from so that I can delete it from that peer? (my understanding is that the sdclient log files only show the core and preferred servers names and not the peer names even if the files get downloaded from the peers). I cannot use the xtracing as this client has already downloaded the files and I can't try it on a different computer as I can't reproduce it easily.
2) Why would the peer have the wrong file?
3) Why doesn't LANDesk try to download a file with a mismatch hash from a different preferred server or from the core?
Attached is the log for this particular client.
sdclient_task235.log 15.1 K