-
The first thing that occurs is the list of URLs in the cache is scanned. The
classification of the URLs (normal, offline, stray, etc.) is done later.
-
The next thing that is done is files in the cache subfolders are scanned and
compared against the URLs already scanned. If there are no matching URLs for a
particular file, it is marked as a stray file. If a URL doesn't end up with a
matching file, the URL is marked as a stray URL. Note that if you do not have
strays set to be scanned every time a regular URL scan is done, neither stray
files nor URLs will be found on passes where the stray scan is not done.
-
All of the scanned entries are sorted by last access time, so the default
deletion order will be "least recently used".
-
The content in the cache is classified by type and a corresponding manager is
assigned to each file/URL. Since it is possible for a file/URL to match
multiple content managers, the managers at the top of the manager list get
assigned first if a match is found. The list is then traversed manager by
manager until a match to that content is found or the end of the list is
reached. If no manager matches the content, no manager is assigned and that
content will never be deleted. (You generally do not want this to happen; a
catch-all manager should always exist even if it is set to not delete content
for a long time.) If the content is assigned a manager that deletes it
immediately, it will do so if the deletion wait period has passed. (See below)
Otherwise, if the manager is set to include this content's size in the cache
size calculations it is added to the size of all the content in the cache. This
is repeated for every entry (normal, cookie, and stray) in the cache.
-
If the content in the cache is larger than the maximum set size, the cache is
reduced in size until it is down to that maximum size. The managers at the top
of the list delete their content first, followed by the managers lower in the
list until the cache is reduced adequately. Since a manager may have a deletion
wait time set, not all of the content a particular manager handles may be
deleted until that content has aged enough.
-
After all of the above has taken place, CacheSentry Pro goes back to waiting or
sleeping until the next management cycle occurs.
| THE SETTINGS PAGE
PROPERTIES |
Overall cache size
The first setting you will see on this page is the overall
cache size setting. This is where you set how large you wish to have the
Internet Explorer cache grow. The value is is megabytes and this setting is
global - it applies to all of the content in the cache as a whole and not to
individual cache managers. Note that this value is not a hard limit - depending
on how you set the managers to wait for files to age a certain amount before
they are available for deletion, the cache may grow larger than the value you
set here.
Manager list selector
Next you will see a control that lists and lets you select the manager you want
to edit. Do a left mouse button click on the down-pointing arrow to see the
list. You can scroll through the list with the scrollbar located on the right.
When you select a new manager from the list, its properties are updates in the
buttons and controls displayed below. Also, the selector lets you type into it
so you can change the name of the currently selected manager.
Change Order
When you click this button, a window will pop open that will allow you to
change the order in which the managers are run during a cache management cycle.
Managers at the top of the list are run first, while managers at the bottom of
the list are run last. To change the order, select a manager in this window you
wish to move and use the up and down arrows on the left hand side to move it up
and down in the list. Press OK when you have finished changing the order of the
managers. If you change your mind and want to undo your changes, press Cancel.
Add Manager
Clicking this button creates a new manager in the list. It will be given a
generic name so you'll want to rename it to something meaningful to what it
manages. In general, there is no limit to the number of managers you can have
in a set but remember: The more managers you have, the more work (and time) the
CPU will take to manage the cache!
Remove Selected
Removed the currently selected manager from the list. Once you delete it, you
cannot get it back so you will have to recreate it if you make a mistake. It is
OK to have no managers in the set but of course that is totally pointless! :-)
Load List
Loads a new set of managers from a file on your hard drive. A dialog window
will open so you can select the CacheSentry Manager set (.csm) file to load.
The current list of managers will be replaced by the list in the new set. Also,
this new set will become the set that will load automatically the next time you
start CacheSentry Pro.
Save List
Saves the current set of managers to a file. A dialog will open allowing you to
select the name of the file to save the list under. This command also makes the
newly saved set the default set which will load the next time CacheSentry Pro
starts.
MANAGER PROPERTIES
This section is the set of properties/settings that define how the currently
selected manager behaves.
Type of Content
Sets the type of content this manager can work on. There are several types:
-
Normal: This is the content you see when you visit a web page by typing
in URLs or clicking on links. If you select this type for this manager, this
manager will only see normal web content.
-
Offline: This is content that is downloaded on a schedule (or manually
with IE's "synchronize" command) that you have set up to be viewed offline.
When you create a favorite for a web site, you may have noticed IE gives you
the options of making it available offline. Content that you have made
available offline is seen by managers with this type set.
-
Normal and offline: A combo of the above; the manager will see both
normal content and offline content.
-
Stray files: A manager will only see stray files if it is set to this
type. A stray file
is a file that exists in the cache but does not have a URL record associated
with it. (All of the URLs are stored in a hidden file named index.dat.) Some
stray files are not actually stray; they are part of a file that has not
finished downloading completely. This can be anything from a web graphic to an
executable program.
-
Normal and stray files: A combo of the above; the manager will see both
normal content and stray files.
-
Normal, offline, and stray files: The catch-all of the cache. The manager will
see all three of the above types of content in the cache.
-
Stray URLs: The manager will only work on stray URLs. A stray URL is the
opposite of a stray file - there is a URL record in the index.dat file, but the
actual physical file associated with it is missing.
-
Cookies: Select this for a manager that will only manage cookies.
Manager sees files named
Here you can enter a pattern to match files against, and the manager will only
work on files that match this pattern. You can enable this filter (and enter
the pattern) by checkmarking the box on the left of it. There are some
important rules to know about when using pattern matching:
-
The pattern matcher is simple in order to keep it speedy. You can specify a
wildcard with the usual "*" (asterisk) character, but it can only appear in one
location in the pattern. For example http://www.enigmatic*
is legal and will match any URL that begins with "http://www.enigmatic". Also,
something like http://www*.jpg is also legal
and will match all entries that start with "http://www" and end with the ".jpg"
in the filename/URL name. However, doing something like *enigmatic*
will not work.
-
You must remember to use a pattern that makes sense for the type of content you
are working with! If you are working with URLs (from normal, offline, or stray
URLs) you must remember to include the "http://" in the pattern string if your
wildcard is at the end (or you aren't using a wildcard for this pattern),
otherwise it will not see the URLs in most cases. (Since most URLs start with
"http://".) Simply specifying only the web address (www.someplace.net for
example) won't work. The same is true for cookies. All cookies start with
"Cookie:", so don't forget to include that in your cookie managers that use
pattern matching. Stray files also have a special consideration: They don't
match against anything except the filename since they don't have URLs.Also, IE
usually appends bracketed numbers to the filename before the extension (for
example "picture[1].gif") so you will have to remember this if you plan to do
any pattern matching with stray files.
-
If you enable the pattern matching but leave an empty string, that is the same
as leaving the pattern matcher disabled. In other words, the manager will not
be filtered by naming and will see all content that matches its type and the
other filters. Also, specifying a single "*" with no text in the pattern also
will match to all names. (However, please remember it is more efficient
CPU-wise if you simply disable pattern matching if you plan to match with
everything.)
-
Pattern matching is case-insensitive. This means it doesn't matter whether you
specify the pattern in upper or lower case, or some combination, as it will
match against both upper and lower case names. For example HTTP://WWW.HI.COM
is the same as http://www.hi.com
to the program.
-
If you really need to match patterns that have multiple names you can always
create multiple managers that handle each part of the name.
Files this manager sees...
This setting allows you to limit the manager to only see content that is
larger, smaller, or equal to a size (in kilobytes) that you set. In order to
enable this filter you must checkmark the box on the left of it.
This manager will not delete its files until...
This prevents a manager from deleting its assigned content until it hasn't been
used for a set amount of time. By "unused" that means the last-access time
hasn't changed for that specified time period. The main purpose of this setting
is to prevent the manager from trying to delete content that might still be in
use. (Trying to delete content that is still in use results in an error #32
message in the activity log.) Also, you can use this setting to ensure cached
content stays around for a minimum amount of time if you want to make sure
certain types of content are always available in the cache. (For example the
favorites icons.) You can specify the amount of time a file must go unused for
in either days or hours. When you change one value, the other one is
automatically calculated. The hours value is always the value that is used -
the days value simply calculates the hours value for you.
Size calculation inclusion
This setting determines when and if the content that is assigned to this
manager is included in the total cache size calculation. You may ask why you
would ever not want to immediately include the size of content in the cache
size calculation, and the reason is simple: If you download a large file it
will very likely wipe out a large portion of your cached content, even while it
is downloading! Also, content that you may want to always exist in the cache
can have its inclusion in the cache size calculation disabled so it appears
"invisible" and won't affect the deletion of other content from the cache.
Conversely, content that you deem as truly "temporary" (content that will get
deleted as soon as it is seen) you probably won't want to include in the size
calculation as that will protect all of the content that you do want
to keep from being deleted before this "temporary' content. There are three
ways to control size inclusion:
-
Content is not deleted until it passes its aging value set in the "When to
delete..." setting described above. Obviously, if you don't use the "when to
delete" setting for this manager, this option will not be available and will be
grayed out.
-
Always include content in the size calculation. This is the main setting to use
for standard content.
-
Never include content in the size calculation. Pretty much self-explanatory.
Stray URLs and cookies always use this setting (and it is not changeable)
because cookies are separate from the cache (and thus don't belong in the cache
size calculation) and stray URLs don't have any real files so they don't take
up any real space. (They take up space in the index.dat file, but because the
index.dat file is not included in the cache size calculation, stray URLs
effectively don't take up any space from CacheSentry Pro's point of view.)
Delete content ASAP
Normally, you want the cache to fill to the maximum size you set in the maximum
cache size setting and do not want content to be removed until that limit is
reached. However for some nuisance content you may want to delete such
content immediately. For example, the zero-byte stray files generated by
Outlook Express are deleted immediately (after they have gone unused for a
while because they do nothing other than take up space in the hard drive's
directory listing (and slow the filesystem down). Note that like the size
calculation, you cannot set this for cookies and stray URLs because again those
two types of content aren't included in the cache size calculation and
therefore waiting until the cache fills up is meaningless to them. (They are
always deleted as soon as any set "aging" has passed.)
-
Manager order is very important, and there is somewhat of an art to getting the
order just right. See the section below for some insight as to why I put the
managers in the included sets in the order they are in. If you can't figure out
why your particular manager is seeing/deleting content before another manager
even though it is lower in the list, post your manager set to the message forum
and there are users and support people than should be able to help you.
-
When thinking about manager order, remember that you may want to put a
particular type of content high on the list of managers because you either want
to 1) delete it before anything else, OR 2) save
it from deletion! Once content gets assigned to a manager, it doesn't get
re-assigned to a lower listed manager when it also matches that manager's
settings.
-
Experiment! Because CacheSentry Pro has the "simulate" button, you can try out
different cache manager settings without worrying about accidentally deleting
all the files in your cache. Plus you can also back up the cache if you are
still nervous about a new manager you have created.
-
Try not to go crazy with the pattern matching feature by creating many, many,
pattern matched managers since the pattern matcher requires the most CPU work
out of all the manager filters.
-
Don't forget to disable the Auto-manage button when testing managers! Even
though the program automatically turns off the auto-manager while you are in
the Size Settings page, if you go to the Activity Page to see the results of a
simulation it will turn on again and you could lose a lot of content if there
is a mistake in how you have set your managers. (When the automanager is
re-enabled, it waits 5 seconds before doing a cache management cycle.)
-
Again, make sure all content is handled eventually by something. If you allow a
group of content to go unmanaged, it will go unnoticed and keep growing in size
in the cache until the cache reaches the Internet Explorer
cache max size setting. And once that happens, any content can be deleted from
the cache!
-
Use the activity log to see how old the last listed piece of content being
deleted is during a management cycle. This is a great way to tune your
overall cache size to a size that keeps the content around for a predetermined
amount of time. This way, you don't have to go through all of your managers and
individually set the "no deletion" aging time. Use the "no deletion" age
setting for keeping files in use from being deleted, or when you want to
strictly force content to stay in the cache for a period of time.
| ABOUT THE INCLUDED
MANAGER SETS |
The factory default set
-
Zero-byte strays: This manager runs first because zero-byte strays are the most
useless of all the content types. :-) Actually, because these files are deleted
immediately (after a 2-day wait to make sure they aren't in use - this is done
in case the zero-byte file is actually a 500MB file someone is downloading over
a slow connection. :-) ) and not when the cache is full it really doesn't
matter where this manager appears in the list! If there are any files assigned
to this manager they get deleted immediately in the cache sizing part of the
cycle.
-
Large files: Next to get deleted before anything else are large files that
might eat up a large portion of the cache very quickly. Usually these will be
downloaded programs but if you view a lot of large videos and music files on
the web you may want to remove this manager or fine tune it. It uses the
powerful "don't include in overall size until deletion wait time passes"
feature. This means it won't affect other cached content because as soon as it
ages enough it gets deleted before any other content since this is the second
manager in the list. Once it ages, its size gets included and if the size goes
over the max (since it has the "cache must be full" option set) bye-bye large
file! Note that we include all types of content in the large file category:
Normal, offline and stray.
-
Favorites icons: This is high in the list because we want to save
this content from deletion! If we put it lower in the list below the more
generic managers they would see this content and delete it before we wanted. By
assigning icons to a manager early (and one that keeps it around for a good
while) this content is unavailable for assignment to other managers and
therefore can't be deleted by those lower-level managers. Note that this
manager also looks for and saved stray icons - this is because stray icons seem
to still be able to be used as favicons in certain cases.
-
Stray URLs: This could also appear anywhere in the list since it is always
deleted immediately during the sizing part of the cycle, but it's simply placed
here in the order to make things less confusing. Again we wait a bit before
deleting them to make sure they aren't in use.
-
Stray files: These make up true stray files and incomplete downloaded content.
Again they must age by 2 days before being deleted. If you have a high speed
connection you may want to lower this value. If you only plan to manually run
the management cycle you could even set this to zero if you don't use IE while
you're using CacheSentry Pro. Unlike the original CacheSentry, this manager is
set to not delete the strays until the cache is full just in case a file really
is part of an incomplete web page load.
-
Normal content: This is the catch-all manager. It sees both normal and offline
content (because we handled both types of strays above, and as you'll see
cookies below) and deletes it with no special conditions when the cache is
full. If you plan to use a really
small cache size (like 1 or 2 megabytes) you may want to enable the deletion
age wait feature to something like 1 hour in case the content is still in use
and thus avoid the annoying "file in use" error messages in the activity log.
-
Old cookies: The last manager in the list. Since cookies are always immediately
deleted when conditions are met, again it doesn't really matter where managers
that handle cookies are in the list. (Unless of course you have multiple cookie
managers, in which case you will have to figure out how to order them in
relation to one another.) Again it's at the bottom of the list just to make
things less confusing and separate it from the other managers.
David's default manager set
This is what I use on my own computer. It is basically the same with a few
additions to handle my development environment needs:
-
MSDN help files: For some reason, the MSDN help system stores everything you
view in the cache, even though it's already coming straight from the hard
drive!
Because I don't want those files pushing out content that came from the
Internet, it is the second manager in the list (right below zero-byte strays)
and matches both normal and offline content and uses the name match feature to
assign to it only files that start with "ms-it" in the URL.
-
Local web server: I set up an Apache server on Linux in order to test my web
pages. Because I really don't want content from this speedy LAN-based server
pushing content out from the Internet, it also gets deleted before the other
web content. It is next in the list, and again uses the pattern matching to
match against all local IP addresses on my network.
-
Maya doc server: Like the local web server, I don't want content that is
quickly available cluttering the cache, so any URL that starts with
"http://localhost:" is removed from the cache quickly. (This should also work
for any local server on your system that uses the same type of URL.)
-
MSDN Platform SDK .cab files: These files come from the MSDN Platform SDK site,
and they are quite large - so large that it can day several days (for me :-))
to download them! So I set up an exception to the "large files" manager and
content from the Platform SDK is spared from deletion for a lengthy time and is
ignored from the actual cache size so they don't push out other content.
 |
Cool idea: If you use the
Platform SDK site and want to keep a local copy of the installation files, you
can use the CacheSentry Pro backup tool's name filter to back up just the .cab
files. They can then be deleted until the next time you need to install another
Platform SDK. This way, you can restore them to the cache, the SDK IE installer
applet will see them and and you won't have to wait for all the dependent .cabs
to download again. A major savings of time and bandwidth!! |
Time-based manager set
This set of managers is an example of how to force content to stay in the cache
for a specific amount of time. Zero-byte strays are still deleted ASAP, but
large content, stray files, and normal content all must go unused (and
therefore stay around) for at least 20 days before they can be considered for
deletion. Icons, stray URLs and cookies are still the same as with the default
manager set. It doesn't really matter how small you set your cache size to -
the usable content will always stay around for at least 20 days. Plus, it may
stay around longer if the cache hasn't filled up to that maximum cache size
setting. If you want to use this set and have content forced to stay around
longer, go through each of the managers that is set to make their content not
be deleted for 20 days and increase the value. (Or lower it if you want to
shorten the content lifespan.)
Manual stray cleaner
This set demonstrates how to set up CacheSentry Pro to be a manual
stray file cleaner. It cleans out all of the stray files immediately. Just make
sure you turn auto-management off otherwise you'll lose all of those files as
soon as the first management cycle occurs! This manager set contains no other
managers except for the stray cleaner, so it won't touch any other files.
Manual cookie cleaner
This set immediately cleans out your cookie folder. Again, make sure
auto-management is disabled. It is basically the same as the stray cleaner
above, except it looks for cookies instead of stray files for the content type.
Return
to the Table of Contents/Main help page