Re: Mysql sleeping queries

[David Reynolds]

Malcolm,

On 19 Jun 2007, at 8:09 am, Malcolm Tredinnick wrote:

It would be interesting to know if anything prior to [5482] (that's
5482, not 5492) works. There was a very large change in [5482] that

aware of.

should be mostly invisible but may have some side-effects we need to be

Thanks for your advice - I went back to 5481, which seems to have cleared it up.

The details, if you care: template output is now generated via

around. This should be mostly invisible. However, one thing that might
be observable is that functions which are called to generate the

to the client. Ticket #4612, for example, shows how this can be a bit


Now, I have no idea if this is affecting you, but off the top of my
head, that is the biggest change we have pushed out lately. I can't
think of any MySQL specific changes, but I'm getting old and my memory
is failing sometimes, so maybe I've forgotten something.

iterators, rather than constructing a huge string and passing the string template may not necessarily be called until the data is being sent back subtle in middleware (not actually a bug, just some experience gained).

Thanks for your help,

David
--
David Reynolds
[EMAIL PROTECTED]

I didn't mention that we were using mysql_old (although I guess that was perhaps obvious from the version of MySQLdb). Having fixed the sites we have running trunk (by going back to r5481) we have noticed that the few sites we still have running pre-magic-removal also have sleeping processes - is this normal? Is there anything we can do to stop this?

smime.p7s
Description: S/MIME cryptographic signature