[Zope-DB] Re: Reviving an old thread :: DCO2 and memleaks...

Geir Bækholt Geir Bækholt
Fri, 10 Jan 2003 10:10:29 +0100


Hello Dario, 
Thursday, January 9, 2003, 4:06:32 PM, you wrote:

> ----- Original Message -----
> From: "Geir Bækholt" <geirh@funcom.com>
>> Did anyone ever find a workaround or fix for this ?

> Nope, not yet.

> We have no real clue of what the problems might be on the zope part of the
> system. We do have a lot of oracle connection objects and we are trying to
> cut down on them down to one per oracle schema (don't ask why we had more
> than one per schema previously).

Our SQL is all fine, but i have several Oracle connections (to the
same schema) within the same site structure. Are there indications
that this may cause trouble by itself?

>...try to limit the amount of CPU-time that oracle will give to each query, and
> if it is exceeded we abort the query. This is to avoid hangups on the db
> server. Hopefully this will also give us info on what queries are the ones
> affected.

This seems more like a workaround than a fix to me. We have monitored
the hung connections, and cannot seem to find any trouble with the
calls themselves. The SQLmethods return all fine, and everything seems
fine from within Zope, the connections (well, most but not all)
just do not terminate...

To me, it looks like a bug in DCOracle2, but i have absolutely no idea
where to start looking..

> I will report back on whether this is sucessfull or not.

Great , thanks ! :)

-- 
Geir Bækholt        geir@funcom.com
Tools/HCI-developer
Tools/Billing - Product Operations
Funcom Oslo