Patches freezing / not exiting cleanly

[almaris]

I have been receiving numerous reports of patches not exiting cleanly on many machines. I have even seen the problem myself several times. The only way to get past it is to quit the task via taskmgr or right click the task in the taskbar and exit. This does not affect the patch, as it seems to have successfully finished at this point; but it is however a problem I consider to be very serious. Does anyone have any experience with this issue? If so, what steps have you taken to fix it?

Thanks!

[jassing]

Do your log files show anything?
do you run any post-patch processes that might be having issues?

[almaris]

Unfortunately the logging is not enabled for patches, but as for post-patch processing, there is an optional launch of an executable but the freeze occurs before the finish screen is shown (after the patch application is finished)

[jassing]

Unfortunately, I think your 1st step is to enable logging.
and add a LOT of logging to each step/process so you can capture exactly where it hangs.

[almaris]

Just a minor update for closure's sake,
The issue ended up being a font we were attempting to install during the post-patch process that did not exist in the patch file list

[jassing]

Glad it was somethign simple (even if it was hard to find)
Thanks for letting us know.

[margre]

We have a lot of problems that the patch stops updating the progress after a while. Windows will add "not responding" to the titlebar. It seems to be worse for large files.
Customers will kill the patch because they think that the patch has stopped working. That destroys the installation since the patch is writing to the target file while patching (we are not using the rollback support). The cheksum for such a file will not match any version, so they have to reinstall.

[jassing]

Generally speaking, it's not good to hijack a thread, especially one that's been resolved...

In your case - The program probably isn't "locked up" and is very busy patching away; if you just leave it alone, it would, no dobut, return and your system would be fine. Lots of programs get that annoying "stopped responding" tacked to their title, especially when they are tight in a loop or processing large amounts of data.

The trade off for this tight code/intense processing is that it appears to "hang"; what you get for this is faster patches. If the program introduces "delays" so that a progress bar can nicely update or to prevent windows from thinking it's "stopped responding" the processing (ie: patching) would be significantly slower.

If you can give us a sample that 100% hangs, we can certainly look at it in more detail; but here; however, I think impatience is to blame.

Best to start a new thread for a new issue.