Is this the same behavior for a session with no parallel processes as well?
________________________________
From: oracle-l-bounce@xxxxxxxxxxxxx <oracle-l-bounce@xxxxxxxxxxxxx> on behalf
of Hemant K Chitale <hemantkchitale@xxxxxxxxx>
Sent: Thursday, June 15, 2017 5:01 AM
To: mwf@xxxxxxxx
Cc: ORACLE-L; jeff.d.smith@xxxxxxxxxx
Subject: RE: tabs, connections, and resulting sessions from sqldeveloper
Mark,
You "hold and reserve" those PX Server processes until you either (a) fetch
*all* the rows or (b) run another, preferably, non-PQ statement.
I tell me developers to run select 'x' from dual if they aren't fetching all
the rows.
Hemant K Chitale
Sent from my smartphone
On 15 Jun 2017 05:14, "Mark W. Farnham" <mwf@xxxxxxxx<mailto:mwf@xxxxxxxx>>
wrote:
Thanks my friend.
One side issue, which I just want to make sure I understand correctly
specifically for sqldev: When you actually fire a query and get back the first
50 row (default) and then do nothing on that screen, is it correct if perchance
the query was fired at, say, parallel(32), then those parallel server producers
and consumers will still be there waiting for the user to take some action such
as getting the next “50” or ending the query.
Other than by timeout, I don’t see how Oracle could close those “servers” at
least until they have queued up all the rows that eventuate, and possibly not
even then.
If this is a figment of my imagination, that is fine. Otherwise, though, I’ll
need to explain to some folks that opening a couple dozen tabs as private
connections, firing a like number of parallel queries and then walking away for
an hour or two is not a peachy keen idea.
TIA,
mwf
From: oracle-l-bounce@xxxxxxxxxxxxx<mailto:oracle-l-bounce@xxxxxxxxxxxxx>
[mailto:oracle-l-bounce@xxxxxxxxxxxxx<mailto:oracle-l-bounce@xxxxxxxxxxxxx>] On
Behalf Of Jeff Smith
Sent: Wednesday, June 14, 2017 3:09 PM
To: mwf@xxxxxxxx<mailto:mwf@xxxxxxxx>; ORACLE-L
Subject: RE: tabs, connections, and resulting sessions from sqldeveloper
Each connection results in 2 actual connections.
One for reports, the browser/tree, and the worksheet.
A second one for bg queries used to power things like the Insight feature in
the SQL Workheet.
If you open an ‘unshared’ worksheet to a database – it opens a private
connection for just that worksheet that the rest of the product doesn’t have
access to. I’m not sure if unshared worksheets also get the 2nd connection for
insight feature. But with unshared worksheets, each additional worksheet would
result in an additional connection.
From: Mark W. Farnham [mailto:mwf@xxxxxxxx]
Sent: Wednesday, June 14, 2017 9:04 AM
To: ORACLE-L <oracle-l@xxxxxxxxxxxxx<mailto:oracle-l@xxxxxxxxxxxxx>>
Subject: tabs, connections, and resulting sessions from sqldeveloper
Anyone, but Jeff it you’re lurking in particular: Is there a simple document
summarizing the relationship of database connections, tabs, running queries,
and so forth from sqldeveloper to the number of sessions created and held?
(Hint: I’m looking for something shorter than reading the entire concepts
guide. The target readership is impatient, would be well-served by reading the
entire concepts document, but that is unlikely to happen.)
mwf
