Solaris page cache limit

Hello all,
on a Solaris 10 (7/08) system we have here, we see the page cache increase over time. Does anyone know until what level this will increase and how to force it not to exceed a certain threshold?
thanks,
Twan

It will gradually use all memory except for a meg or so as page cache.
Thats what solaris does, if theres any memory thats not being used for actual work, it figures it might as well use it as page cache.
But the page cache will be released immediately if anything else wants the memory.
So you can think of the page cache pretty much as free memory. If fact most solaris tool simply report the page cache as part of the free memory.
So why would you want to limit the page cache?

Similar Messages

Solaris 8 cyclical page cache

I have read Sun's article about Solaris 8 cyclical page cache, but still don't understand how it works.
It did not mention how does kernel devide memory between the IO page buffer cache and applications.
Is it is first come, first served basis?
Is there any way to size the IO page buffer cache?
When app needs a page, and there are none any available, does it take it from IO buffer cache?
Does it mean that IO page cache will be 0 in size if application requires all physical memory?
In short, what is the algorithm?
I see me database application swap the executable/data parts heavily. It looks like it is due to kernel page cache competing for memory with the app. So, the problem is still not fixed after introduction of "priority paging" in Solaris 2.7 and "cyclical page cache" in Solaris 8?
Thanks for any feedback.

It will gradually use all memory except for a meg or so as page cache.
Thats what solaris does, if theres any memory thats not being used for actual work, it figures it might as well use it as page cache.
But the page cache will be released immediately if anything else wants the memory.
So you can think of the page cache pretty much as free memory. If fact most solaris tool simply report the page cache as part of the free memory.
So why would you want to limit the page cache?

Report (not Page) caching to improve report loading times

We are trying out Cystal Reports Server 2008 as a replacement for Crystal Reports CR XI R2 for an ASP.NET Web Application. Running some tests we found that the same report loaded from the server was far slower than loading the same report locally, using the CRXI SDK.
Being the Clone/Refresh strategy on the standalone SDK several times faster than the reportAppFactory.OpenDocument on the server platform, i think that the server is loading from scratch the report each time I execute the OpenDocument method. So what I am seraching for is, Is there a way to setup or tweak report caching on Crystal Server 2008? What I need is not page caching, but the whole report, which is used different databases each time it's executed (implying different data for each execution).
Some insight on the problem:
As I posted on a previous thread we are moving from CR XI due to a limit on the number of active instances suported by the SDK (74 instances). We reach this limit because of a workaround for improving report execution times, leaving instances of each report active, so the next execution just recicles the active instance, greatly reducing the overall process time. Therefor introducing Crystal Server is a setback on this module (due to it's actual performance).
Part of the problem resides on the complexity of the report. Using the standalone SDK the report takes several seconds to load from disk, and several seconds to execute. Regretably, we are not at liberty to change the reports' structure, so optimizing it is beyond posibility for now.
Thanks in advance,
Gustavo

Hello Maggie,
>> How can I get an all-encoompassing CSV file of the main report (Page 2014)?
This might be possible using the advanced print server configuration, with BI Publisher, using the same technique that is being used to print master-details reports (which is a type of a multi-region report) - http://www.oracle.com/technology/products/database/application_express/html/configure_printing.html . The standard print server configuration only supports reports with a single region. If you have BIP in your organization, that’s great. Otherwise, CSV files don’t warrant it.
The only other option, I can see, is to create the CSV file manually, using the technique described in the following Blog entry, by Scott Spendolini - http://spendolini.blogspot.com/2006/04/custom-export-to-csv.html .
Regards,
Arie.

Clear the page cache

Hi,
We have recently started to develop our application on Solaris, before each regression test is run we need to clear the page cache so that performance is consistent. On Redhat/CentOS there is the following command:
To free pagecache, dentries and inodes:
echo 3 > /proc/sys/vm/drop_caches
However I cant find an equivalent command for Solaris. We have been using tail -f /dev/zero to run a huge process until the memory has been flushed which isn't very reliable and slow. Anyone know of a different way to do this?

If we are running multiple tests one after the other then the first test may make calls which causes information to be cached, then when the second test runs it may make use of the previously used pages in cache. This makes the second test results inconsistent as if it ran first the timing results would be different, also running through a debugger or with truss doesn't give the accuracy for timing that is required in this situation.

Clear page cache every time page is opened

I have a menu page in Apex that points to 10 report pages. However each page retains the values in parameters. How can I clear the page cache every time I go from menu to individual report page? I see the CLEAR_PAGE_CACHE procedure but don't know where to place the call. I guess I haven't understood the event model of the Apex completely yet (I am from the Oracle Forms background).
Jay

Jay,
You don't say how these links on the menu page are constructed. If you are using an apex list then for each list entry there would be a target page section in which you would put the target page ID in the Clear Cache field. For example, if the link goes to a page in the current application, say page 10, then you'd put 10 in the Clear Cache field.
If you are building a link yourself, just put 10 (for page 10) in the clear-cache position of the f?p URL, e.g.,
f?p=&APP_ID.:10:&APP_SESSION.::NO:10:P10_ITEM:&SOME_ITEM.
Scott

Apex 4.2 200 Page Items limit

Hi,
I have a master-detail form which was getting close to the 100 items per page limit in Apex 4.0 which I then scaled back and got to work properly. I had thought that I could start adding more columns to the detail part of the form since 4.2 was supposed to have increased that limit. I am finding that this does not seem to be true at least in the case of this particular form. I tried increasing and started getting 404 errors when trying to save data--same symptom I had had with 4.0. I am back nearly to the same # of columns I started with--and nowhere near the 200 limit. Columns are month by month projection data that I had hoped to extend with the increased page item limit.
Am I missing something required to take advantage of larger page items limit? Do I need to use one of the new templates?
Any pointers would be appreciated!!
Pat

Hi Pat,
changing the compatibility mode will have no affect in your case. First I would check the real reason for the HTTP 404 error. See http://www.inside-oracle-apex.com/oracle-apex-got-404-not-found-2/
I assume it's not the page item limit you are hitting, instead I think you are running into the 2,000 parameter limit of mod_plsql, because your tabular form has too many rows/columns. See mod_plsql:/pls/apex/wwv_flow.accept HTTP-400 too many arguments passed
Regards
Patrick
My Blog: http://www.inside-oracle-apex.com
APEX Plug-Ins: http://apex.oracle.com/plugins
Twitter: http://www.twitter.com/patrickwolf

FAQ: Are there best practices for building projects within the 20 page/state limit?

The 20 page/state limit in Flash Catalyst is there to prevent Catalyst (and your finished application) from running slowly. You can however build efficient applications that have more than 20 states by using custom components. Custom components can contain states as well; so by creating an app that has several states, and using custom components that have states, you can get more unique views of your app while keeping it efficiently built.
Try this:
1. Select some of your arwork where you need more states. Right click and choose "Convert to Custom Component"
2. Double-click to edit the custom component. Note that you can now create states in the custom component. Try creating a few states here.
3. To exit editing the component, double click a blank area on the artboard.
4. If you try creating an "On click transition to state" interaction now you will see that you can choose from both the states of the application as well as the states of the custom component.
Answered by: Ty Voliter. See entire discussion.
More help:
Video tutorial on custom components
Video/demo discussing the benefits of "pushing interactivity down" into custom components, by Ian Giblin @ MAX 2009
(jump to the 11:15 mark, and watch through to about 20:30)
Another forum post by Ty:
Here's an ASCII diagram to illustrate one way of pushing application or top-level pages/states down into a custom component:
Before:
State1     State2     State3     State4     State5     State6     State7...
After (project refactored to push three top-level states down into a custom component):
State1     State2     State3      State7...
                    |
Custom Component
State4     State5     State6
For some projects this can be an inconvenient workaround, and for others it can really clean things up by reducing complexity. If you have several top-level pages in your app that are completely different than other pages (for example, they don't share any objects with the other states) than pushing these down into a custom component makes things a lot more manageable in Catalyst. The layers panel, for example, shows all objects across all states-- this is useful when your states share many objects. If however, your states don't share a lot of objects, the layers panel can get a bit unmanageable. Refactoring into a custom component allows you to fix this by grouping content into a container.

The 20 page/state limit in Flash Catalyst is there to prevent Catalyst (and your finished application) from running slowly. You can however build efficient applications that have more than 20 states by using custom components. Custom components can contain states as well; so by creating an app that has several states, and using custom components that have states, you can get more unique views of your app while keeping it efficiently built.
Try this:
1. Select some of your arwork where you need more states. Right click and choose "Convert to Custom Component"
2. Double-click to edit the custom component. Note that you can now create states in the custom component. Try creating a few states here.
3. To exit editing the component, double click a blank area on the artboard.
4. If you try creating an "On click transition to state" interaction now you will see that you can choose from both the states of the application as well as the states of the custom component.
Answered by: Ty Voliter. See entire discussion.
More help:
Video tutorial on custom components
Video/demo discussing the benefits of "pushing interactivity down" into custom components, by Ian Giblin @ MAX 2009
(jump to the 11:15 mark, and watch through to about 20:30)
Another forum post by Ty:
Here's an ASCII diagram to illustrate one way of pushing application or top-level pages/states down into a custom component:
Before:
State1     State2     State3     State4     State5     State6     State7...
After (project refactored to push three top-level states down into a custom component):
State1     State2     State3      State7...
                    |
Custom Component
State4     State5     State6
For some projects this can be an inconvenient workaround, and for others it can really clean things up by reducing complexity. If you have several top-level pages in your app that are completely different than other pages (for example, they don't share any objects with the other states) than pushing these down into a custom component makes things a lot more manageable in Catalyst. The layers panel, for example, shows all objects across all states-- this is useful when your states share many objects. If however, your states don't share a lot of objects, the layers panel can get a bit unmanageable. Refactoring into a custom component allows you to fix this by grouping content into a container.

Do shared photo streams have a cache limit or revolving cache?

I stopped using shared photo streams a few months ago because I noticed that the photos were stored locally on my phone and in iCloud. When I checked my storage usage in settings I noticed that 500mb+ was taken up by the shared photo stream storage. I have since started using Flickr, but miss the convenience of shared photo streams to create quick albums.
Does anyone know if there is a revolving cache that limits the data that shared photo streams store on a device? If shared photo streams start to replace its own data after 500mb I'd be happy to give up that space for convenience.
Does anyone know if there is a cache limit to shared photo streams, and if so, what that limit might be, and if it is revolving?

All photos that you want to share to a stream have to be kept on your device as well for syncing with iCloud. There are limits to how many photos you can share to the shared streams, however. See:
iCloud: Photo Stream limits

Page Zero (0) - Regions Caching / Page Caching not available

Hi,
My page zero display information from a calendar table (fiscal year, etc.). This information will be the same all day long (get updated at midnight). It's not possible to enable caching of regions on page zero. It's also impossible to enable caching of the page zero.
Why?
Thanks,
Louis-Guillaume
[My Blog|http://lgcarrier.blogspot.com]
[APEX and jQuery demos|http://www.insum.ca/jquery]

Louis-Guillaume,
Good question. We have asked ourselves this and are considering removing the restriction in 4.0. I think it initially had to do with preventing unwanted side effects and anomalies although none of us can recall the details right now. We will have to carefully look at ways to make this something developers can use easily while preventing strange results. First of all, caching of page 0 doesn't make sense. You want to be able to cache regions on page 0. Now, say your page 0 has two regions, P0_CACHED and P0_DYN (one cached, one dynamic). And your page 10 has region P10_CACHED. When page 10 is rendered, you'll get:
P0_CACHED
P0_DYN
P10_CACHED
That looks okay, you get one cached region from page 0, one dynamic region from page 0, and one cached region from page 10. Of course you have controls with which to purge any individual region from the cache so you can cause P0_CACHED to be refreshed whenever you like.
But say page 20 is a cached page. When it renders you'll get:
P0_CACHED <-- not from the page 0 region cache but from the page 20 page cache, regardless of the "stale-ness" of the page 0 cache
P0_DYN <-- not rendered dynamically from page 0 but retrieved as part of the page 20 page cache. But this content may be different from the content for the same page 0 region displayed on page 10 one second ago or one second from now.
P20_DYN <-- from the page cache as is normal for a dynamic region on a cached page
Hardly a thorough treatment, I realize. Just wanted you to know some of the aspects we have to consider. We will also have to read this thread again to come back up to speed: V3 Caching - any more info? .
Scott

Problem with page cache clearing

I have a button that exits HTML DB and returns to our portal main page. That is done through a URL, and not a page number. There is no way to reset page cache with a URL branch. So I made a page that has a branch that calls the URL, and then I changed the button on the page to branch to this page and do the proper page cache resets before calling the other page that branches to the URL. However, the page cache is not being reset when I do this.
Here is the page flow
page 10 - Select a record for update
page 20 - Display data using an ID that is on page 10
If Return to Menu is clicked branch to page 1 and reset page 10 and page 20
Page 1 - Branches to URL
When I go back to page 10 from the application server page page 10 & 20 are still populated.
If I change page 20's return to menu to branch to page 10, then the reset does occur. What am I doing wrong? How do I reset the pages when leaving HTML DB?

Dwayne,
Validations, like other components, can have conditions that observe the current request value. This value is set to the button name when a button submits the page. So you could have a condition on a validation like 'When request != Expression 1' and in the Expression 1 field type the button name (check the button attributes to confirm the case-sensitive name and submit the page in debug mode to see the value of request when you click the button).
About clearing page cache when branching to URL -- not sure what you're after: Cache can be cleared if an HTML DB page is invoked via URL. You specify clear-cache options in the branch definition which you can observe in the URL that invokes the target page. If you mean that you want the cache of one or more HTML DB pages to be cleared "when" a branch to a non-HTML DB page is taken, that doesn't really make sense because nothing much happens during a branch except to call the HTML DB engine's accept procedure or to do a URL redirect. So all you can do is specify some action to happen "before" a branch is taken, e.g., clear cache on pages x,y,z using a page process, or specify that the clear cache action is to occur "after" the branch is taken as manifested in the resulting URL to an HTML DB page. For something to happen "when" a branch is being done presumably would be after page processing but before a redirect, which we already support in the form of page processes.
Scott

Image from BLOB now showed when page cache set

Hello,
I have a page that display images from BLOB,
After setting page cache to 1hour, page load faster but images not showed.
Please help me on this.
Thank you.

Hi Halit,
For me i dont thing, this problem occurred. As i checked by setting the page cache to 1 hour. After that images loaded normally as usual.
I tried with my mozilla browser. In which browser you tried, may be your browser fault.
Kindly check it.
Brgds,
Mini
Mark Answers Promptly

Partial Page Caching in 902

What, if Any, partial page caching using ESI tags can be done in Portal V2 or with the PDK?

There is no PPC available through Portal in 902.
PPC will be available in 903/4.
Obviously you can still use PPC in your on code that directly accesses Webcache
Thanks
-JP

Putting a page cache as old

Hi, we are caching page definition & content, but at sometimes
we want to force a recalculation programmatically, i.e., to put
the page cache as old.
I tried to used the wwutl_api_cache package, but it (silently)
doesn't change anything. We are using Portal 3.0.9.8.
Any clues?
Thanks in advance.

> When I press F12 to preview the site, the link to the
Orient page is as
> follows:
>
> C:\Documents and Settings\Tracy\My
Documents\Serenity\Orient
> But how do I make it so the .htm is there and doesn't
need to be typed in
> order to preview the page? I have this problem with two
different pages.
okay-- so the path is wrong.
Open the page for editing in dreamweaver.
Select that link with the mouse.
In the Properties Inspector, use the folder icon next to the
Link box to
browse to the correct file.
Alan
Adobe Community Expert, dreamweaver
http://www.adobe.com/communities/experts/

How to disable a JSP page caching?

          Hello there,
          How do I can disable a JSP page caching? This is becoz when the first time i ran
          a EJB calling from a JSP, it works, but after the second time, it unable to call
          the same EJB again, I double check the console and the database, the same records
          does not appears after the second calling, i suspect is caching problem.
          TQ
          Neo
          KL


Neo, wls does not cache jsp page by default. We do have a custom tag for
          it as you can see here
          http://e-docs.bea.com/wls/docs61/jsp/customtags.html#overview.
          Regards,
          Joseph Nguyen
          BEA Support
          "Neo" <[email protected]> wrote in message
          news:3d4a428d$[email protected]..
          >
          > Hello there,
          >
          > How do I can disable a JSP page caching? This is becoz when the first time
          i ran
          > a EJB calling from a JSP, it works, but after the second time, it unable
          to call
          > the same EJB again, I double check the console and the database, the same
          records
          > does not appears after the second calling, i suspect is caching problem.
          >
          > TQ
          >
          > Neo
          > KL

Invalidating portal page cache

On a portal page, contents are rendered by portlets using java programs. With page caching on, the entire portal page is cached. Performance is great when users are navigating back to this portal page from other pages.
Because the portlets hold session context in the application, this portal page is invalid once a user signs out or closes the browser window. The problem is that if this portal page is cached, when the user signs in again portal would present the cached page without going to the portlets to re-establish a new session context.
What can we do to invalidate the cache of this portal page? We found nothing in the portal to specify any invalidation rules for a portal page. We believe we cannot invalidate the cached portal page from the application.
Any ideas will be appreciated,
Chung

With the following settings:
1. Portal page caching:
Cache Page Definition Only
2. Provider.xml of portlets:
<pageExpires>20</pageExpires>
3. Provider registration info:
Require portal user specific session information. (Option checked)
we get the following results:
- When portal user navigates back to the portal page consisting the portlets,
the page is rendered quickly. It's because provider.xml says portlets are good for 20 minutes.
- When a user logs out and within 20 minutes logs in again IN A NEW BROWSER WINDOW, a new session id will be used, portal will not use the cached portlets. Hence, requests will be made to the application and a new application session context will be created.
However, if the user uses the same browser window, then cached portlets will be used. It would encounter the application session timeout error. This restriction might be acceptable to customers.
We will recommend the above settings to the customer. Please verify that this is a valid solution.

Solaris page cache limit

Similar Messages

Maybe you are looking for