B
Brian
We're using a Sybase 8.0.3 "SqlAnywhere" database. And it's twitchy.
If anyone has an idea how to clear the train wreck, I'm all ears. What
happens is that even the slightest ODBC error (say an implicit cast) will
hang the current cursor and all subsequent cursors indefinitely (okay, maybe
two hours or so). The website becomes useless not only to the person who
originated the bug, but to everyone who tries to invoke ODBC calls later.
Everyone sees the browser icon spin.
Here's what I've done so far:
I set anything that looks like a timeout value to 10 seconds. No effect.
Set a few ansi parameters per bug reports.
Set a timeout thread to abort the call. This worked on the immediate page.
But it still left the train wreck. All other users were still stuck.
Updated to the most current version of Sybase 8.
Theoretically, if everything is perfectly programmed, it shouldn't be a
problem. But as has already been the case, another programmer brought the
site down with a simple bug (not closing the connection). What a mess.
On a similar note, is On_Page_Unload guaranteed to run? There's no common
bug or asynchronous event that prevents it from cleaning up, right?
Thanks for reading. Hopefully someone can help me out!
If anyone has an idea how to clear the train wreck, I'm all ears. What
happens is that even the slightest ODBC error (say an implicit cast) will
hang the current cursor and all subsequent cursors indefinitely (okay, maybe
two hours or so). The website becomes useless not only to the person who
originated the bug, but to everyone who tries to invoke ODBC calls later.
Everyone sees the browser icon spin.
Here's what I've done so far:
I set anything that looks like a timeout value to 10 seconds. No effect.
Set a few ansi parameters per bug reports.
Set a timeout thread to abort the call. This worked on the immediate page.
But it still left the train wreck. All other users were still stuck.
Updated to the most current version of Sybase 8.
Theoretically, if everything is perfectly programmed, it shouldn't be a
problem. But as has already been the case, another programmer brought the
site down with a simple bug (not closing the connection). What a mess.
On a similar note, is On_Page_Unload guaranteed to run? There's no common
bug or asynchronous event that prevents it from cleaning up, right?
Thanks for reading. Hopefully someone can help me out!