I have an SSIS job that has been running on our test server since April, with no issues. Last week it just quit working. About the fourth package in I started getting “The script files failed to load” error when loading a very simple VB Script inside a package.
A web search led me to this forum page:
To save you from wading through all the messages, the two causes of this error could be 1) Package’s PreCompile was set to False, or 2) I had breakpoints in the package.
So I check, and nope my PreCompile was set to true (the default) and there were no breakpoints. In addition, most of the people posting in the forum link above couldn’t get their packages to run in the first place, this didn’t seem to occur in packages like mine that had been running fine for a while.
Next I thought, OK what about environmental issues? Mining our extensive logs I was able to determine the time when one moment the scripts work, then 15 minutes later they no longer worked.
With that exact 15 minute interval known, the DBA team reviewed what had happened on the server, and sure enough a security patch for the .Net 1.1 Framework had been applied on the server during that exact time. Bingo!
It took us 2 days to work through the issue, the team, and I have to be honest and include myself here, had the thought process “well, the error is just on the test server, we’ll get to it when we can.” Since it was just a test server, and we weren’t testing anything at the time, we weren’t stressing over it.
We should have been. The very same patch was scheduled to be rolled to our production server just an hour and a half after we found it. Had that scheduled deployement occurred we would have been in deep diaper filling.
Let me make a quick distinction here, in that it’s not necessarily fixing the issue that’s urgent, but identifiying it. Once you have it identified, you can control it. By not quickly identifying the issue, we nearly rolled the same patch into production.
So the first point here, if your SSIS package which has been running for a while suddenly starts getting “The script files failed to load” error, check to see if any .Net Framework 1.1 patches were applied.
Second, this shows you how important good logging can be. With our extensive logging I was able to determine with a fifteen minute window when the problem occurred.
Finally, never treat issues that occur on your test server as if they are not important. We were quite lucky, narrowly avoiding disaster on our production server. Identify the cause quickly, so you can control it.
Update: We found a fix that worked, see https://arcanecode.wordpress.com/2007/07/18/ssis-issue-from-monday/ for more detail.