Alex Karasulu wrote:
On Jan 30, 2008 8:00 AM, Emmanuel Lecharny <[EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>> wrote:
o Entries. They are stored into the MasterTable. An entry is a
serialized Attribute, associated with a unique identifier, which is
currently a Long (which allows our server to store up to 18 446
744 073
709 551 615 entries ... 18 quintillions !).
Yeah but you had an idea that was great today. The id today is
partition specific: meaning it is unique only within a partition. As
you suggested we need to make sure that it is unique to the server.
This will make it so we can conduct searches better which traverse
across multiple partitions.
And we need to do this because we intend to nest partitions and do
away with the PartitionNexus.
To summarize : having a global ID generator induce that we have one
place (synchronized) to ask for a new Id. The main issue with this
approach is that we must persist this Id on disk when we stop the server
- this is obvious - but also we should be able to ressucitate the 'next'
Id wrom the disk if we have a crash (remember that we don't write data
on disk immediatly, as we have a differed write mechanism. Not
necessarily a problem, though, as we can store this id when writing the
data on disk every N seconds (dependening on the sync timeframe)
Those long must be fetched
quickly, as we will always get an entry by its ID. Using a BTree for
that is time consuming, as fetching an entry will be done in
O(log(N)).
You're absolutely right a hash would be much better. We don't need to
sort the ID's.
Howard has a very valid point regarding the use of Hashtable vs BTree.
o Attributes with single values. Typically, 'uid'. There are
supposed to
be unique, too (ie, you can imagine that an uid can be found more than
once in a LDAP server (there is no reason why you should not be
allowed
to store a user's UID in many places, if this user is registred in
more
than once...
There is no uniqueness constraint on UID in LDAP. Perhaps you mean the
entry UUID attribute?
No. I was mentioning the SINGLE-VALUE vs the MULTI-VALUE. Uid is
single-valued.
For instance, I use the same uid on several computers, so I
may have to store these UID in different branches in my LDAP
server). We
can use a Hash table to store such an attribute, because it's supposed
to be unique. If you have more than one reference, then the question
will be : how do I manage more than one value associated with this
attribute (cf next point).
Using a Hash table here means you know that this attribute is not only
holding a single value, but also it's unique within the whole server.
Dangerous ...
I think this is not a good idea for UID but it could be great for a
UUID index.
If we were to use a HashTable, then I would say : this is up to the user
to determine if he want to use a HashTable or a Btree for an attribute's
index. this should certainly not be a static choice we make.
o All other attributes (including single valued attributes which
can be
found more than once) : HashMap is not really a good choice, except if
the admin *knows* that the number of values won't be enormous.
This is not a good idea because you need sorted access and traversal.
Think about an index on age. You have the > and < operators where you
have to advance to some position and walk the index finding all
entries that comply.
In fact, the index are depending on the operation. You can use a single
index for < or > filters, but you need another one for *=, and another
one for substring operation. I have described some of those in the badly
outdated page I pointed.
<IDEA>
Incidentally that just gave me an idea. The server's search mechanism
is a bit dumb where ranges are given. For example if I give it this
range (age>45) and an index exists on the 'age' attribute then a
cursor is walked and all entries on the walk are returned. So there
is a partial walk of the index. Now when I give it this filter (&
(age>45) (age<60)) then say a cursor is chosen for the first
subfilter. That walk happens and the second is used to assert that
the age is less than 60. When the cursor gets to 61 for example, it
should stop but it does not - the walk continues but now the second
subfilter says nope don't return it because it's over 61. We could
have in this case just abandoned the walk but the server does not
because it's too stupid. Some smarts are required to merge numeric
ranges into a special kind of filter node.
Yeah, that's true. You need to stop a traversal as soon as you reach the
upper limit. We have to think about how to handle this...
</IDEA>
The current implementation, as explained by
Alex, instead of storing a list of all the entry IDs linked to the
'person' value, stores a TreeSet, which is managed in a separate file.
Not in a separate file. The TreeSet is persisted as the value of a
multivalued key to implement duplicate keys. Careful not to mix this
up with attributes or multivalued attributes.
Note that after some threshold though we switch to using another BTree
to redirect to. So when we do a lookup and find that the value is a
reference (BTreeRedirect) then we just use another BTree in the same
RecordManager (same file).
That what I meant. You have an indirection, and it costs time.I thought
the redirected tree was stored in another tree, I'm glad this is not the
case. The important point is that the redirected BTree is not totally
loaded in memory.
There is an indirection, a little bit like when you deal with a N-N
relation in a RDBMS context (when two tables are related with a N-N
relation, then you create a third table holding the relation between
both IDs).
I don't get this analogy at all. It's totally off.
It's not. Your redirectBTree is this third table. Think about it twice,
you will see the analogy.
Can we do better ? Because this solution is costly, and not really
effective : you have to deal with the structure duality (you hold
either
a list or a reference to a treeset, depending on the number of
elements), and this make it more complex to manage cache (i'm not sure
at this point that we can cache such elements...)
It's cached by the record manager. The record manager caches blocks
in the db file with data not the records themselves.
Good point. BPages are loaded in cache, and holds the elements. The way
it works is not really our business... I have mixed two different caches
here. Forget about what I said.
Now, for those attributes, I have described another solution, based on
the fact that we use 'long' to store IDs :
http://cwiki.apache.org/confluence/display/DIRxSRVx11/Backend. I think
this might be interesting to dig those ideas a little bit to see if it
fits our need.
Yeah there were some neat ideas you had. They're pretty experimental
so we need to test them out. Would be nice if we can get our lab up
and running to test the differences in design.
BTW this backend documentation incorrect when it explains the present
implementation. Please yank that page and just extract your ideas to
expose them because otherwise people learn incorrect things when you
publish documentation with misconceptions in them.
Ok, I got the message :) This page is pretty old and contains bad
assumptions about the way the backend is implemented. I will update it
today (at least, I will mark some parts as outdated)
I apologize if I sound negative - I rewrote the above 3 times to make
it softer :).
Thanks buddy ! It does not sound negative, it's just a fact : it's
outdated, and misleading !
--
--
cordialement, regards,
Emmanuel Lécharny
www.iktek.com
directory.apache.org
