More or less recently this "DownloadAltPatchDataSource" procedure started to fail due to the Kaseya 2hrs timeout on procedures.
This procedure is written by Kaseya and the Agent Procedure Logs are showing, at the moment of the fail, it was attempting to download a file called WSUSSCN2.cab (URL: http://go.microsoft.com/fwlink/?LinkID=74689) from Microsoft using CURL.exe.
But here is the hitch: I believe that CURL.exe is partly responsible for the issue...
When manually initiating the same download using CURL.exe,
it seems to be going fine at the beginning: the file size grows as expected.
However, the CURL.exe
process does not end (still running in the task manager)... and after a few minutes, the size of the downloaded
file get reinitialised to zero, and the download starts again... and again...
It seems as though CURL.exe
does not wait long enough to let the download complete (my guess: a hard coded
timeout inside CURL.exe).
It looks like the timeout
period is too short, and CURL.exe recycles the download from scratch.
When attempting to manually the download the file Internet Explorer, the same sort of behaviour happend: stalls at 99%.
However, after waiting for 5 to 10 minutes, the download is in the end successful.
A Ticket is currently open regarding this matter: CS088418 .
I however am curious to find out if any of you have experienced the same issue...
Thanks for you replies...
Hi Francis - yes, we're experiencing the same thing. Typically only on clients with slower internet connections. It's blowing data caps and causing frustration.
Greig @ IT Partners.
Francis - did you get a resolution? We're still seeing it intermittently, and if your ticket has been resolved, it would be great to know what the answer was.
After numerous tests on our K server, the K-Team came out with a change in the DownloadAltPatchDataSource procedure.
In short, the command line used to call CURL.exe (or CURL-NOSSL.exe) is getting new switches.
They prepared a Hotfix and installed it manually on our Kserver.
I've just sent then an update telling them I haven't had any procedure failure notification for a few days now. Since the procedure runs twice a day (12:30 am/pm over a 6 hrs distribution period), I told them it was looking very promissing.
Currently, the Ticket has a "Hotfix Review" status.
I am confident they will release the hotfix quite soon.
Many thanks Francis. I appreciate it.
FYI, this hotfix has been released (applies to 6.2)
Here is the Hotfix info (taken from our Hotfix History):
17-Mar-12 6.2.0 Hotfix CS088418 - 1887 - Changed dynamically created agent procedure used to download the WSUSSCN2.CAB file to the LAN share to use a direct call to cUrl.exe rather than calling the GetUrl() function call to resolve intermittent download failures in low bandwidth environments MessageSys\Hotfix\PatchProcess.pdb
17-Mar-12 6.2.0 Hotfix CS088418 - 1886 - Changed dynamically created agent procedure used to download the WSUSSCN2.CAB file to the LAN share to use a direct call to cUrl.exe rather than calling the GetUrl() function call to resolve intermittent download failures in low bandwidth environments MessageSys\Hotfix\PatchProcess.dll