Aw: Re: Re: Re: [ApacheDS] Porting 1.x based custom partition to 2.0
Having this example comes in very handy, thanks for that. But I have two questions: is it already foreseeable when the interface will be fixed ? Reason for my question is that I want to build our new solution on a final release and not on a milestone. Second question: is there an example how to implement a partition with a custom backend ? I know there were examples for JDBC in 1.x but did not find one for 2.0 | Gesendet: Samstag, 14. September 2013 um 15:16 Uhr | Von: Kiran Ayyagari kayyag...@apache.org | An: users@directory.apache.org | Betreff: Re: Re: Re: [ApacheDS] Porting 1.x based custom partition to 2.0 | Here[1] is the updated sample using M15 (M16 has changed the way a | partition is initialized, so | used M15 instead to keep it clear for the sake of understanding, except | that there is no major change | in initializing with M16) | | ... | [1] | http://svn.apache.org/repos/asf/directory/sandbox/kayyagari/embedded-sample-trunk
Aw: Re: [ApacheDS] Porting 1.x based custom partition to 2.0
3. do you know of any examples for this for 2.0 ? I only know examples for 1.x and know that these won't work for 2.0 I guess not, I will update the example code to work with the latest trunk and let you know I do not want to rush anyone so please take this as what I indend it to be: a polite question. Did you already find the time to update or can you estimate when this will be available ? Regards,
Aw: Re: Re: [ApacheDS] Porting 1.x based custom partition to 2.0
I know where the problem is but haven't committed anything yet, I can commit it in the next week only That would be great and I am looking forward to it.
[ApacheDS] Porting 1.x based custom partition to 2.0
Over a year ago I developed a custom partition with ApacheDS that I want to port to 2.0 now. The partition is special because it is only used to connect to ApacheDS, the search in the sub tree is then taken care of in Javacode and not in a directory or database. The result for the query needs to be determined live without copying any data to ApacheDS beforehand. This is not my exact scenario but I want to give you an example which might explain what I want to do. Let's assume I want to use ApacheDS to look up square roots. It would not make sense to store all possible results in ApacheDS. A search for a square root would look like this: ldapsearch -b ou=squareroot -M -h localhost -p 389 -x (value=25) squareroot will be the name of the subtree / partition I write. The search is value=25 which tells the partition that I want to receive the square root of 25. The partition would then simply calculate Math.sqrt(doubleTakenFromQuery) and return the value. So my questions are: 1. do you rate 2.0 as mature enough to build my solution on it ? (I assume yes but perhaps a certain Milestone is a better fit) 2. is it still the best way to use a partition for this or does 2.0 offer better ways ? 3. do you know of any examples for this for 2.0 ? I only know examples for 1.x and know that these won't work for 2.0
Re: [ApacheDS] How to setup a debugging environment
Is this an error on my side is something wrong with the files I downloaded ? You can't run eclipse eclipse before having built the project at least once : the SNAPSHOTs aren't stored into the maven repository. Just run mvn clean install -DskpiTests first, then re-run mvn eclipse:eclipse So i obviously got the order wrong. Now I succeeded with these commands. I set maven opts to the mentioned xmx setting and had to add a maxpermsize parameter, but this might be because of the buggy jdk I had installed (updated afterwards): C: cd \workspaces md apacheds cd apacheds svn co http://svn.apache.org/repos/asf/directory/apacheds/trunk-with-dependencies cd trunk-with-dependencies mvn clean install -DskipTests (Results can be found in trunk-with-dependencies\apacheds\all\target and trunk-with-dependencies\apacheds\service\target) mvn eclipse:eclipse Then went to eclipse and chose File / Import / General / Existing Projects into Workspace. For root directory I chose C:\workspaces\apacheds\trunk-with-dependencies Now I have a snapshot M8, can I do the same for M7 ?
Re: [ApacheDS] how can a partition return more than one result ?
Von meinem iPad gesendet But I am only able to return ONE entry, I didn't find or understand the concept how MULTIPLE entries can be returned. Can someone show me the right direction ? The idea is to use a Cursor that maps around the partition and fetch the entries one by one. The way the server works is that based on your filter, you select the right index to use to fetch the entries. There are may possibilities here : - first, you may have to do a full scan (the filter is not selective enough, for instance). In this case, you don't use any index, you just use the MasterTable to get the entries. Now, for each entry you fetch, you'll have to filter them to see if it's a valid entry - or not. - or you can select an index. You will fetch the index elements, and for each of them, fetch the associated entry. Once done, you can check against the filter if the entry is valid - or not In any case, the cursor is your friend here : it maps the next() operation on top of your index. Now, if your Partition is a Btree, it's easier, as the AbstractBTreePartition class already handles everyting for you. If you don't inherit from this Abstract class, then it's way more complicated. I'll suggest you have a look at the AbstractBTreePartition to get a clue about how we process a search over a BTree based partition. If I got you right I am to create my own cursor object which retrieves and stores the array with the items I want to deliver. The cursor then implements the next and get method in order to send the array one by one. But where is the cursor being called ? I tried to base my own partition on the abstract btree partition but what am I supposed to return from getRootId and getDefaultId and even worse from convertAndInit ? Sorry for the beginner level questions but ApacheDS is quite complex and new to me.
Re: [ApacheDS] How to setup a debugging environment
Am 09.06.2012 um 00:03 schrieb Emmanuel Lécharny elecha...@gmail.com: Le 6/8/12 11:38 PM, garb...@gmx.de a écrit : In order to broaden my understanding I want to connect Eclipse to a running instance of ApacheDS 2.0M7. I retrieved the sources for M7 from the svn repository but failed in combining them. I do NOT want to build ApacheDS, my only intention is to make all the subprojects known to Eclipse to allow it to remotely debug ApacheDS. I want to connect to a running instance and use the debugger to trace what's going one in partitions. Can you explain me how to setup a debugging or if necessary a build environment in Eclipse ? You can run 'mvn eclipse:eclipse' to produce all the .classpath and .projects for Ads. I'm afraid you have to first build the project itself, which takes quite some time, unless you run 'mvn clean install -DskipTests' to avoir running the tests. I have no maven infrastructure yet so I will start from scratch. Is it ok to use version 3 ?
[ApacheDS] how can a partition return more than one result ?
I learned a lot about the implementation of custom partitions and see the changes in the API from search and lookup returning an Entry in 1.5.x and an EntryFilteringCursor later on. I even was able to change an existing partition (shame on me: based on the 1.5.5 example, but I will switch to 2.0 soon) in a way that lets it return a fake group object that was created in my custom POJO. But I am only able to return ONE entry, I didn't find or understand the concept how MULTIPLE entries can be returned. Can someone show me the right direction ?
Re: [ApacheDS] how can a partition return more than one result ?
Am 08.06.2012 um 13:58 schrieb Emmanuel Lécharny elecha...@gmail.com: Le 6/8/12 1:49 PM, Garbage a écrit : I learned a lot about the implementation of custom partitions and see the changes in the API from search and lookup returning an Entry in 1.5.x and an EntryFilteringCursor later on. I even was able to change an existing partition (shame on me: based on the 1.5.5 example, but I will switch to 2.0 soon) in a way that lets it return a fake group object that was created in my custom POJO. But I am only able to return ONE entry, I didn't find or understand the concept how MULTIPLE entries can be returned. Can someone show me the right direction ? The idea is to use a Cursor that maps around the partition and fetch the entries one by one. The way the server works is that based on your filter, you select the right index to use to fetch the entries. There are may possibilities here : - first, you may have to do a full scan (the filter is not selective enough, for instance). In this case, you don't use any index, you just use the MasterTable to get the entries. Now, for each entry you fetch, you'll have to filter them to see if it's a valid entry - or not. - or you can select an index. You will fetch the index elements, and for each of them, fetch the associated entry. Once done, you can check against the filter if the entry is valid - or not In any case, the cursor is your friend here : it maps the next() operation on top of your index. Now, if your Partition is a Btree, it's easier, as the AbstractBTreePartition class already handles everyting for you. If you don't inherit from this Abstract class, then it's way more complicated. I'll suggest you have a look at the AbstractBTreePartition to get a clue about how we process a search over a BTree based partition. -- Regards, Cordialement, Emmanuel Lécharny www.iktek.com Thanks, I will investigate this. What a pity, I am able to map searches to string arrays containing the name of groups I want to return. But I understand why you don't support arrays directly, I will somehow manage to map to the Bree example. And if not I will show up here again ;-)
[ApacheDS] How to setup a debugging environment
In order to broaden my understanding I want to connect Eclipse to a running instance of ApacheDS 2.0M7. I retrieved the sources for M7 from the svn repository but failed in combining them. I do NOT want to build ApacheDS, my only intention is to make all the subprojects known to Eclipse to allow it to remotely debug ApacheDS. I want to connect to a running instance and use the debugger to trace what's going one in partitions. Can you explain me how to setup a debugging or if necessary a build environment in Eclipse ? -- NEU: FreePhone 3-fach-Flat mit kostenlosem Smartphone! Jetzt informieren: http://mobile.1und1.de/?ac=OM.PW.PW003K20328T7073a
Re: [ApacheDS] lost with decision where to implement custom permission
Am 07.06.2012 um 17:59 schrieb Emmanuel Lécharny elecha...@gmail.com: Le 6/6/12 6:39 AM, Garbage a écrit 1. Can M7 be considered stable enough for giving it a chance in a productive system? No. I'm sad to say that but M7 still have a huge bug that we are tracking. Not that you'll won't get a working server, but in some cases where concurrent writes and searches are done, you are very likely to get some failures during the searches. More than that, the database can get corrupted. We exactly know what is going wrong here, and we already have a fix, but this is in a branch, and we are currenly merging this branch into trunk, which is all but easy. M8 will solve this issue, and we expect to get it released ASAP. In the mean time, you can use the directory/apacheds/branches/apacheds-txns branch whch solve the issue. I only need ApacheDS to sit there and wait for incoming connections and forward them to my custom partition. There is no need for replication and all the other fancy stuff ApacheDS is able to do. You can still work on the trunk, until M8 get released. Just keep tuned, we are doing our best to get back to some solid and reliable server soon. I believe I can use the released milestone 7 because parallel reads / writes don't exist in my scenario. Or are there any other issues I should keep an eye on ?
[ApacheDS] Question regarding caching behavior
From what I know ApacheDS supports caching of search results, that means when I issue the same search after e.g. one minute the result will be returned from the cache. First question: is this correct ? Is this something ApacheDS does on it's own or is this the job of the partitions involved ? Second question: when implementing a custom partition would I need to take care of caching on my own ?
[ApacheDS] lost with decision where to implement custom permission
I tried to build a working custom partition implementation for last days. I started with 1.5.7 but then found out that the working example I intended to base on does no longer compile because of some API changes. Because of this I switched to 1.5.5 and managed to get the example running (not that difficult, I have to admit ;-)) Unfortunately I don't have enough background knowledge to be able to extend the HelloWorldPartition example. I'm willing to learn and invest time but need expert advice if I should continue with 1.5 or 2.0. In addition I will need some guidance. In exchange for that I can offer you to update the docs for custom partitions.
Re: [ApacheDS] lost with decision where to implement custom partition
Sorry, I either mistyped or became a victim of auto correction: I meant PARTITION Am 05.06.2012 um 18:28 schrieb Garbage garb...@gmx.de: I tried to build a working custom partition implementation for last days. I started with 1.5.7 but then found out that the working example I intended to base on does no longer compile because of some API changes. Because of this I switched to 1.5.5 and managed to get the example running (not that difficult, I have to admit ;-)) Unfortunately I don't have enough background knowledge to be able to extend the HelloWorldPartition example. I'm willing to learn and invest time but need expert advice if I should continue with 1.5 or 2.0. In addition I will need some guidance. In exchange for that I can offer you to update the docs for custom partitions.
Re: [ApacheDS] lost with decision where to implement custom permission
Am 06.06.2012 um 05:16 schrieb Emmanuel Lécharny elecha...@gmail.com: Definitively go for 2.0. The API has changed a lot, but this is (hopefully) for the best. Sadly, the documentation is not following the code modification pace :/ Do you have any specific question ? Thanks for that information, this might be the explanation why I got lost in the transition from 1.5.5 to 2.0: I simply didn't understand what changes were to see. So I will head for 2.0, I expected that answer ;-) But I still have these questions: 1. Can M7 be considered stable enough for giving it a chance in a productive system? I only need ApacheDS to sit there and wait for incoming connections and forward them to my custom partition. There is no need for replication and all the other fancy stuff ApacheDS is able to do. 2. Where should I start off in order to create a partition that uses data from my java objects and sends it back to the client ? Some details to explain what I'm after: I need to integrate a google search appliance with at least three different directories. One is based on LDAP, one can be accessed via a web service and the other has a SQL interface. I discussed this topic on stackoverflow and even posted a question on the developer list of ApacheDS, back then I intended to use an interceptor but now I believe that a custom partition will be a much better solution. The easiest way to integrate these directories will be a LDAP interface because the GSA is able to connect to LDAP, the queries can be configured. The main idea is to write a partition that listens on access to ou=user and answers to queries that follow this pattern: (username=johndoe). It should then deliver the two attributes distinguishedName and mail. In a second step the GSA will send a query that looks like this: (member=distinguishedName). The expected answer will then be a list of dns of groups the user belongs to. This information will be delivered by my java code and can be considered a blackbox. Just think of it returning a string array of groups in ldap format.
