Executing Sql - whats happening on the server

Hi, I am using Oracle 8.17 on a AIX box.
When i run queries that have to do full table scans of a large table it virtually cripples my system. (nmon shows lots of disk activity / paging) .
Even when running
Select count(col1) from large_tab;
Im trying to understand whats happening behind the scenes here. Why is there lots of paging. what is oracle doing?
Our database is set to rule based optimiser as the vendor providing the software does not allow for use of cost base optimizer / or compute statistics. :-0 :-(
Why would this be do you think?
thanks for any advice you could provide me.

have you set minperc & maxperc aix parameter ?No , please can you let me know what these are for? I
cant seem to find much on the web relating to them?sorry :
minfree / maxfree
you can monitor this param with vmtune
minperm maxperm minpgahead maxpgahead minfree maxfree pd_npages maxrandwrt
157079 628316 2 8 120 128 524288 0
One simple basic setting for minfree and maxfree is to calculate a starting point as follows:
" minfree = ( 120 * # of CPUs )
" maxfree to ( ( maxpageahead * # of CPUs ) + minfree )
on metalink you read this values :)
others request ?
sga ram swap ?
felice_pago

Similar Messages

What happens to the server if I put a System.exit(0) in my servlet...

deployed and running it?
M

Normally, my Servlet container uses whatever JVM I have installed. For example, Tomcat does not normally come bundled with a special, Servlet-based JVM. So, when I use 1.4.2 (or whatever), System.exit() should be honored immediately. Or do you believe that the ClassLoader detects when a Servlet has loaded and alters the behavior of the JVM?
Now, maybe there is some bizarre command-line switch you can specify in your JVM startup command to cause it not to honor System.exit(), but other than that possibility, it seems to me that Java is Java. If System.exit() halts my application, it should also halt a Servlet container, which, in turn, is also a Java application.
- Saish

What happens when the multiply effect is applied to both the object and the graphic within?

Hi everyone,
I could really use some help with this...
I'm getting a booklet ready to be printed. Originally, half of the pages were to be printed on paper with a (physical) colour. For budget reasons however, I've now decided to print the paper colour, and have the content of the pages overprinted on the tint in the background, which I've applied to the pages in the background of an additional master page.
When I set the objects to multiply, they blend with the background colour, which is the effect I'm looking for, as there would be no 'paper white' on the pages that I want to appear as though printed on coloured paper.
There's something confusing me though; you can also set the graphic within the object frame to multiply. This makes the graphic appear darker, which I don't really understand, since there is no extra colour within the graphic that accounts for this. I've looked at the output in the resulting pdf, and it looks like it's not just an on-screen effect, but the extra amount of (in this case) yellow would also be visible in the printed result.
Does anyone have an explanation for this? Please see the graphic for a visual example.
Thanks in advance, kind regards
Robbin

Hi.
          Yes. There is a heartbeat mechanism between all cluster members so they detect when a
          member has left the cluster. The primary server recognizes when it's secondary server
          has failed, and then proceeds to search for another member of the cluster to make as
          it's new secondary.
          Regards,
          Michael
          David Whitehouse wrote:
          > I'm trying to figure out what happens when the server holding the replicated state
          > for an object fails.
          > Does the server with the primary object (or the replica-aware stub) recognize
          > the failure and select a new
          > server to hold the replicated state?
          >
          > For instance, Assume I have a cluster defined with three servers - 1,2,3 - with
          > an instance of object A on server 1
          > and the replicated state ino on server 2. What happens if server 2 fails? Is the
          > state now replicated to server 3?
          >
          > Thanks, David
          Michael Young
          Developer Relations Engineer
          BEA Support

VDI3: what happens when the Primary VDI core fails?

A potential customer asked me this and I did not really know how to answer the question ... sorry if this was asked before. So here we go:
What happens when and if the Primary VDI core fails? Will the entire VDI3 setup go down? Will running sessions be lost? And how do you fix that?
I found this page:
http://wikis.sun.com/display/VDI3/MySQL+Cluster+Reconfiguration+Scenarios
So I assume the "Non-VDI host ==> Primary Management Host" scenario is what would need to be done?
http://wikis.sun.com/display/VDI3/MySQL+Cluster+Reconfiguration+Scenarios#MySQLClusterReconfigurationScenarios-NonVdiToPrimary
If this is correct then this would answer the "how do you fix that?" part. OK then.
But still I'd welcome if anyone could explain what exactly would happen when and if a Primary VDI core fails. The customer's opinion on this is that they are a "single point of failure" because the Primary VDI host is not redundant (as opposed to the 2 x Secondary VDI cores) ... So, is this assumption correct or not?
Thanks in advance.

Hi,
Assuming you have 3 VDI core hosts. One of them is the primary. And the primary goes down. So what happens then:
* The underlying database is still running on the remaining hosts
* All desktop sessions are still running on the remaining hosts
* New session requests will be handled by the remaining hosts.
* All desktops are still running on the virtualization hosts.
So in essence, your VDI cluster is still healthy. The operation is just impacted in this way:
* You can't add new VDI core hosts
* You can't change the configuration of the Sun Ray server failover group
* A failure of another VDI core host (data node) will result into a complete outage of the underlying database
You should bring up the primary again as soon as possible in order to gain failover capabilities again.
-Dirk

What happens when the CC trial ends? What happens when offline?

I need Lightroom to work offline on my laptop. What happens if I need to sign in to CC when there is no interned connection and what happens when the CC trial ends on both my laptop and desktop?

Thanks for answering. I have the standalone version of Lightroom 6 on my Windows 7 64 bit desktop. A number of times when have tried to open the program, the Adobe Application Manager has popped up requiring me to sign in even though I had not knowingly signed out. Lightroom will not launch unless I am signed in. Frequently when Lightroom launches, I see a "Server Busy" message telling me to "switch to the other program" without telling me which program to switch to. I have pinned the AAM to my notification area and have unchecked the "Show OS Notifications box" to see if the "Server Busy" message goes away. So far everything seems to be okay
Again, thanks for answering me

Load balancing - what happens if a server goes down?

Hi, Just an academic question.
Assume I have 2 DA servers behind a BIG-IP load balancer, named DA1 and DA2.
If I enable Load Balancing (BIG-IP in my case), what happens to the clients already Connected to DA1 when IP-HTTPS is used?
For example:
Client1 is connected to BIG-IP, which in turn redirect this client to DA1-Server. Assume that DA1-server goes down, what happens with the client then? I assume the Client will be redirected to DA2-server by BIG-IP after some time, meaning it will have to establish
the connection from scratch. Is this correct and roughly how long does it take for the clients to connect to the DA2-Server?
Since IP-HTTPS is using it's own pool and IPv6 prefix per DA server, I assume this will happen? And the DA client1 will receive a new IPv6 address from DA2 ?
Thanks!

Hi,
It depends of the way your Big-IP is checking the servers in your pool.
The client will try to reconnect to the HTTPS interface and still be redirected to server1 until your Big-IP validate that the server1 is not responding and disable it.
And yes, you'll receive a new IPv6 address because each server in a DirectAccess NLB cluster has a specific IPv6 range for the clients.
Gerald

Value mapping replication - What happens to the cache at restart?

Hi!
Does anybody know what happens with the value mapping infos in the runtime cache of the integration server after a restart of XI?
Are the mapping infos still available or do we need to send mapping values via outbound interface ValueMappingReplicationOut once again?
kr,
Ronald

Hi Ronald,
as far as i know, the cache is being emptied, so you'll need to send the values again.
Unfortunately i can't find any reference to this in any of the documentation, but you can easily test that by replicating data, checking it with the cache monitoring than restarting XI and checking it again.
Regards
Christine

HT201317 What happened to the ADVANCED option on iCloud?

Does anyone know what happened to the ADVANCED option on iCloud.com? I cannot delete my photo stream since Apple has removed that option and replaced it with NOTIFICATIONS.

Jerome.Serpenti wrote:
@Winston Churchill: I hope you will not consider this a breach of your privacy, it was not so intended; but I had to smile at the 'bathroom Apple TV'. I never considered an AppleTV might serve a function there!
It's not a problem, indeed it's helpful, I'm aware it looks as though I'm here to argue, I just don't want others to get wrong information, it helps when others give their opinion.
The bathroom TV, is a TV manufactured for a bathroom, it's a sealed unit, it meets all regulations and sits in the wall where it can be easily viewed while soaking in the tub. The Apple TV is also in the wall and is controlled by our phones, it's connected to an amplifier which is situated in the garage which in turn is connected to waterproof speakers in the bathroom ceiling. I did it all myself and am quite pleased with it.

Pipeline functions - what happens under the hood? (10g)

Hi all,
What happens when you call a pipelined function? (As opposed to a plain table-valued function.)
From what I understand, a pipelined function will return rows as they are needed. Let's say I have a really silly example of a pipelined function,
CREATE OR REPLACE PACKAGE BODY TEST_PACKAGE_1
AS
    FUNCTION GET_DATA RETURN JUST_A_TABLE PIPELINED
    IS
    BEGIN
        PIPE ROW (JUST_A_TYPE('HELLO'));
        PIPE ROW (JUST_A_TYPE('WORLD'));
        RETURN;
    END;
END TEST_PACKAGE_1;
/Then let's say I were to SELECT * FROM TABLE(TEST_PACKAGE_1.GET_DATA) WHERE ROWNUM = 1, what happens to the second row? Does the code that returns the second row still get educated? (Obviously it's omitted from the query, but does it still get generated somewhere and just go nowhere?)
If I have a cursor on a PL/SQL block that selects from GET_DATA, will each row only be returned as the cursor references it? In that case, what happens to the rows that don't get selected? If there some resource that need to be closed out (say, just before the return statement), would it still be safe to close it out?
Thanks!
Don

Interesting question, lets try a simple test and see we'll just take your example function and toss in a few dbms_output statements like so:
CREATE OR REPLACE PACKAGE BODY TEST_PACKAGE_1
AS
    FUNCTION GET_DATA RETURN JUST_A_TABLE PIPELINED
    IS
    BEGIN
        dbms_output.put_line('ONE');
        PIPE ROW (JUST_A_TYPE('HELLO'));
        dbms_output.put_line('TWO');
        PIPE ROW (JUST_A_TYPE('WORLD'));
        dbms_output.put_line('THREE');
        RETURN;
    END;
END TEST_PACKAGE_1;Now not forgetting to enable dbms_output call it three times like so:
SELECT * FROM TABLE(TEST_PACKAGE_1.GET_DATA) WHERE ROWNUM = 1;
SELECT * FROM TABLE(TEST_PACKAGE_1.GET_DATA) WHERE ROWNUM <= 2;
SELECT * FROM TABLE(TEST_PACKAGE_1.GET_DATA) WHERE ROWNUM <= 3;What I saw from my test was that with the first call I got 1 row of data from the pipelined function and only the first dbms_output statement was processed. For the second statement I got 2 rows of data from the function and the first two dbms_output statements were processed. For the final call I again got 2 rows of data from the function and all 3 dbms_output statements were processed.
My environment:
Oracle Database 10g Express Edition Release 10.2.0.1.0 - Product
PL/SQL Release 10.2.0.1.0 - Production
"CORE     10.2.0.1.0     Production"
TNS for 32-bit Windows: Version 10.2.0.1.0 - Production
NLSRTL Version 10.2.0.1.0 - Production

What happened to the Xraid?

What happened to the Xraid? its no longer on the apple website? Are they still in production?

oh. well, the promise, might sound promising at the moment, but im sure there's gonna be some compatibility issues with leopard server and the Xserve soon......
o well. I didnt REALLY need a server and a RAID in the first place.

RI - What if Parent is refreshed - What happens to the constraints

Hello
I have a parent-child scenario. What happens if the parent table is completely refreshed? What happens to the child table (that has cascade delete)?
If RI exists, will a complete refresh of parent be possible?
What is the best way to handle this scenario?
Thx!

One of the quickest ways of identifying duplicate records in a table is to enable a unique (or primary key) constraint knowing ahead of time that it will choke on the duplicates, but using the EXCEPTIONS INTO EXCEPTIONS clause which will cause the rowids of the duplicates to be written into an EXCEPTIONS table (created ahead of time with the rdbms\admin\utlexcpt.sql script, for example).
Therefore, imagine a million row table into which a fresh bulk load of another million rows from some external source is to be performed.
You declare the constraint, you disable it, you perform the load, you enable it exceptions into exceptions. You now can clear out the duplicate data quite easily, and then enable the constraint normally. Next load, you disable the constraint all over again and repeat the process.
That's one reason you might want to repeatedly disable/enable a constraint.
Consider another: a unique or primary key constraint, by default, when created, will cause a unique index to be created on the table. Doing any sort of bulk load into a table, even of data you know to be 'clean' of duplicates, will be a lot slower as a result of each insert having to update the index. Therefore, one of the things you commonly do to speed up bulk loads is to disable constraints, therefore removing any associated unique indexes, and therefore saving the time that would normally be spent validating data and updating indexes.
In the old days before they invented deferrable constraints, too, you might well be engaged in a complex data loading process that involves quite a few steps before finishing. If looked at during those steps, you'd say the data violated constraints... but if you just let me get to the end, you'll find that the data all comes good. So you had to be able to disable constraint checking at the start of the process, and re-enable it at the end. A good example of that is Scott's EMP table: it has a manager column and the data in that column refers to the EMPNO column. A manager has to be an employee, after all! So there's a self-referencing foreign key on that MGR column, which means you can't load a record containing a manager number that isn't already an employee. But if you were doing a bulk load of data into that table, starting with it completely empty, that would mean you'd never be able to load any data into it at all: no employees exist to begin with, so there are no managers, so the first employee record that mentions a manager gets rejected. But if only you'd let me do the initial load of data, you'd see that every manager mentioned DOES turn out to be an employee! At the end of the load, it all comes good, and everything is properly related to everything else.
Short of being very careful about the order in which you load such data (presumably, you'd have to start with Mr King, who runs the company, then deal with the first lot of managers, and then you could load the clerks and salesmen), the simplest way of resolving that sort of Catch-22 is to disable the constraint, do the load, and then enable the constraint when all the data is in place and therefore all the relationships between records have come good.
(As I say, since version 8.0 of Oracle, you can also defer the constraint so that it only gets checked at the point you commit all the data, but leaving that subtlety aside, it's still a classic example of where switching constraint checking on and off proves very useful).

When we do a failover to a Replica, what happened to the connected clients? Assuming that the Replica is located in a remote DR site with a different IP configuration.

When we do a failover to a Replica, what happened to the connected clients? Assuming that the Replica is located in a remote DR site with a different IP configuration.
Hi,
I need some guidance here, please.
I have some questions about replication:
When we do a failover to a Replica, what happened to the connected clients? Assuming that the Replica is located in a remote DR site with a different IP configuration.
Is possible to automate the failover process? Just like happens in a cluster.
What changes do I need to do in order to guarantee that my clients reach my replica? For example, I need to do a failover to my Exchange Server Replica and I need to minimize the downtime. The same applies for a DC or a SharePoint Server
  Thanks in advanced.

Hi efebo,
“The same applies for a DC ”
As for replicating virtualized DC (Personally,  I do not suggest to replicate DC , even though it can be replicate ), please refer to following link:
http://technet.microsoft.com/en-us/library/dn250021.aspx
"In short, Exchange does
not support the Hyper-V Replica feature. Exchange has a long history of supporting virtualisation from
Exchange 2003 onwards. It is fully supported to install Exchange
2007,
2010 or
2013 as a virtual machine on Hyper-V, but using the Hyper-V replica feature is not supported.
For details please refer to following link:
http://blogs.technet.com/b/rmilne/archive/2013/07/29/exchange-and-hyper-v-replica-support.aspx
I assuming that the "Remote DR site" and the primary site both have public IPs .
Based on my knowledge you can not access replica site VM  directly after a disaster , maybe you need to rebuild primary site and replicate VM back .
Best Regards
Elton Ji
We
are trying to better understand customer views on social support experience, so your participation in this
interview project would be greatly appreciated if you have time.
Thanks for helping make community forums a great place.

What happens to the open item in bsid table when it is cleared

hi experts,
i have a doubt about open and cleared items of accounting documents. as we know open items of customers are stored in bsid and cleared ones in bsad. For suupose, an open item is closed ina particular period then what happened to the entry in bsid and bsad table. an entry regarding is stored in bsad (cleared),and what happens to the entry in bsid, still it will be there or removed from it.

Hi,
When ever we clear the document(Complete payment) that document will be moved to BSAD table. Other wise it will remain in BSID.
If u are making partial payment then all the partial payment documents along with original document will be stored in BSID till u made the complete payment for that invoice.
Assumption is u have to give all partial payment documents made for particular invoice item while u are making complete payment in F-28 transaction. So that these partail payment documents will also be cleared while making complete payment and moved to BSAD table.
This trasaction can be used for both partial and complete payments.
Irrespective of complete or partial payments made all the documents will remain stored in BKPF and BSEG tables.
BSID and BSAD are index tables to address performance issues with BKPF and BSEG tables.
So all open and cleared documents will be there in BKPF and BSEG tables.
All cleared documents will be there in BSAD table.
Alll open documents will be there in BSID table.
Hope it is clear.
Thanks,
Vinod.

What happens to the mail in my iMac after I delete it from my iPhone? Will it be deleted from my computer as well?

What happens to the mail in my iMac after I delete it from my iPhone? Will it be deleted from my computer as well?

If they are both using the same iCloud ID (assuming you are running your mail from iCloud), then yes, the email will disappear from your Mac after you delete it from your phone and after your phone sycs with iCloud and your Mac syncs with iCloud.
I use a windows phone, but it works the same.

Just upgraded to 5 from 3 - what happened to the page select icon? Can't find it in customize toolbars.

Used to be you could select which page you wished to go back or forward to from the arrow icons on the navigation toolbar. The history menu item only shows 10 sites without having to open history for today. What happened? Is there another icon or a new plug in for this functionality? I used it all the time.

''what happened to the page select icon (Back/Forward)''
The back/forward buttons have been moved to right-hand side of the navigation bar and the drop-down was removed. User right-click or hold click for one second to see the tab history. This is item #5 in my list.
The History menu also has "Open Previous session" and "Open closed windows",
to see more history use the History sidebar ("Ctrl+H") or the History Library List ("Ctrl+Shift+H") and sort or view by "Last visited".
You can make '''Firefox 5.0.1''' look like Firefox 3.6.19, see numbered items 1-10 in the following topic [http://dmcritchie.mvps.org/firefox/firefox-problems.htm#fx4interface Fix Firefox 4.0 toolbar user interface, problems (Make Firefox 5.0.1, or 4.0.1, look like 3.6.*)]. ''Whether or not you make changes, you should be aware of what has changed and what you have to do to use changed or missing features.''
* http://dmcritchie.mvps.org/firefox/firefox-problems.htm#fx4interface

Executing Sql - whats happening on the server

Similar Messages

Maybe you are looking for