, as this tes never failed.
Is it something due to the OSGi framework, or some concurrent issue with
the tests ?
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
for a full web based Studio
o add configuration support for other servers
This list is just limited, again, please add the items you think it's
important to have.
Comments very welcome !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
on before 2.0, as the underlying
database will be impacted, as we will not serialize the UpValue as a
String but as a byte[].
Thoughts ?
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
really very slick.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 8/22/13 11:32 AM, Kiran Ayyagari a écrit :
On Thu, Aug 22, 2013 at 2:56 PM, Emmanuel Lécharny elecha...@gmail.comwrote:
Making the UpValue a byte[] will save us a lot of wasted CPU, and
probably a bit of space on disk, as a String requires 2 bytes per char
to be serialized.
even the byte
the sea and the sun !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 8/23/13 12:01 AM, Quanah Gibson-Mount a écrit :
--On Thursday, August 22, 2013 11:39 PM +0200 Emmanuel Lécharny
elecha...@gmail.com wrote:
1) OpenLDAP is the fastest, with at least 26 000 search/s, and the CPU
wasn't maxed (we reached 85%). The network transfer rate was around
20Mb/s
(something that should not be too complex to
implement).
Thoughts ?
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Otherwise, I just found this :
http://code.google.com/p/concurrentlinkedhashmap/
AL 2.0, so we may use it.
Worth a try, IMO.
Le 9/2/13 7:11 PM, Mohit Srivastava a écrit :
I think we can use EhCache for now. Also modify it for our purposes.
On 2 September 2013 14:32, Emmanuel Lécharny
when we have a lot of aliases, something unlikely to happen.
I will try to think about a better solution.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 9/4/13 1:27 PM, Kiran Ayyagari a écrit :
On Wed, Sep 4, 2013 at 3:42 PM, Emmanuel Lécharny elecha...@gmail.comwrote:
Hi guys,
I just came back from a short week of excellent vacations, and I spent
the last two days looking at the search operation.
Just before I left, I thought
still have some other ways to gain somemore performance :-)
JDBM performances are 3 times behind (6 381/s) but it was 654/s on the
very first tests.
Le 9/4/13 12:12 PM, Emmanuel Lécharny a écrit :
Hi guys,
I just came back from a short week of excellent vacations, and I spent
the last two days
for many different version
V, you just have to encapsulate those two informations into a new key :
K, V. For instance, if the key is 'acme', and you have 3 versions v1,
v2 and v3, the key you want to use will be acme, v1, acme, v2 and
acme, v3.
--
Regards,
Cordialement,
Emmanuel Lécharny
the cacheService we have from it's
EhCache implementation atm, so that the API is not to be odified after
the next release, and that's it. I also suggest to make this
CacheService available in the Partitions, even if it's not used.
Thoughts ?
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 9/9/13 4:46 AM, Kiran Ayyagari a écrit :
On Sun, Sep 8, 2013 at 1:38 PM, Emmanuel Lécharny elecha...@gmail.comwrote:
Hi guys,
we need to use a cache in the partitions, for entries, aliases, and also
for whatever cache the partitins could need (assuming that this can be
configurable
Le 9/9/13 7:43 AM, Emmanuel Lécharny a écrit :
Le 9/9/13 4:46 AM, Kiran Ayyagari a écrit :
I suggest we focus on decoupling the cacheService we have from it's
EhCache implementation atm, so that the API is not to be odified after
the next release, and that's it. I also suggest to make
Le 9/9/13 8:22 AM, Stefan Seelmann a écrit :
On 09/09/2013 07:52 AM, Emmanuel Lécharny wrote:
Le 9/9/13 7:43 AM, Emmanuel Lécharny a écrit :
Le 9/9/13 4:46 AM, Kiran Ayyagari a écrit :
I suggest we focus on decoupling the cacheService we have from it's
EhCache implementation atm, so
Hi guys,
yestrday and today, I was implementing a cache in Mavibot to replace the
WeakReferences we were using previously. The rational was that with the
WeakReferences, we were unable to inject more than 50 000 entries in the
server (at this point, the GC is just going crazy trying to free some
Le 9/10/13 5:40 PM, Emmanuel Lécharny a écrit :
Hi guys,
yestrday and today, I was implementing a cache in Mavibot to replace the
WeakReferences we were using previously. The rational was that with the
WeakReferences, we were unable to inject more than 50 000 entries in the
server
Le 9/11/13 4:44 AM, Kiran Ayyagari a écrit :
On Tue, Sep 10, 2013 at 9:35 PM, Emmanuel Lécharny elecha...@gmail.comwrote:
Le 9/10/13 5:40 PM, Emmanuel Lécharny a écrit :
Hi guys,
yestrday and today, I was implementing a cache in Mavibot to replace the
WeakReferences we were using previously
,
Cordialement,
Emmanuel Lécharny
www.iktek.com
,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 9/14/13 9:45 AM, Kiran Ayyagari a écrit :
On Sat, Sep 14, 2013 at 1:00 PM, Emmanuel Lécharny elecha...@gmail.comwrote:
Hi guys,
/**
* @return The array of stored values.
*/
V[] getValues();
shouldn't this be returning a BTreeV,V? cause we don't support an array
Le 9/16/13 11:41 AM, Kiran Ayyagari a écrit :
On Mon, Sep 16, 2013 at 3:00 PM, Emmanuel Lécharny elecha...@gmail.comwrote:
Le 9/14/13 9:45 AM, Kiran Ayyagari a écrit :
On Sat, Sep 14, 2013 at 1:00 PM, Emmanuel Lécharny elecha...@gmail.com
wrote:
Hi guys,
/**
* @return The array
On 16 sept. 2013, at 11:51, Emmanuel Lécharny elecha...@gmail.com wrote:
Le 9/16/13 11:41 AM, Kiran Ayyagari a écrit :
On Mon, Sep 16, 2013 at 3:00 PM, Emmanuel Lécharny
elecha...@gmail.comwrote:
Le 9/14/13 9:45 AM, Kiran Ayyagari a écrit :
On Sat, Sep 14, 2013 at 1:00 PM, Emmanuel Lécharny
Le 9/16/13 7:59 PM, Kiran Ayyagari a écrit :
On Mon, Sep 16, 2013 at 5:25 PM, Emmanuel Lécharny elecha...@gmail.comwrote:
IMO, I think we should have the getValues() to return an iterator, and
the getValue() method to returns either the single value or the first
value if there are more than
(),
opContext.getFilter().toString(),controls);
}
}
AXEL GROSSE
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
wanted to keep you informed !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
, but the caller will just have to use it as is.
It seems convenient, at least for a very first version, and it can be
used by the in-memory btree too, eliminating the need of many tests in
the implementation (things like : if ( btree.isAllowDuplicates() ).
wdyt ?
--
Regards,
Cordialement,
Emmanuel Lécharny
Le 10/6/13 4:42 PM, Kiran Ayyagari a écrit :
On Sat, Oct 5, 2013 at 1:48 AM, Emmanuel Lécharny elecha...@gmail.comwrote:
Hi,
I'm currently implementing the ValueHodler for the managed btrees, and
I'm coming with some nice solution for the duplicate values. The idea is
to make
.
So far, so good, we are making good progress, and I hope to be done this
week.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
with improved performances for the
server very soon.
More to come !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
calls, 2 per entry )
o MavibotTable.out() : 19% / 10987 calls
That's for 96% of the time spent in the server.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Hi guys,
I have committed my last changes. I think we should cur a 0.2, I will
prepare it soon.
Thanks !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
,
Emmanuel Lécharny
www.iktek.com
abstain
[ ] -1 Do not release Mavibot 1.0.0-M2
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
the array is correctly used with check (length 128).
You are 100% right.
I have fixed the code accordingly to your suggestion !
See http://svn.apache.org/r1536687
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
for the report, Stefan.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
the Mavibot/ADS problem.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
,
Cordialement,
Emmanuel Lécharny
www.iktek.com
fixed some of them, some more fixes are needed. It will be done
tgis week-end.
thanks for your patience !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
is correct. More to come.
Le 11/1/13 3:33 PM, Emmanuel Lécharny a écrit :
Hi,
just a quick heads up on those failurres.
First, I ran the full tests yesteday, with this module being commented,
and everything went fine.
Second, I still have some failing tests in this module, mainly because
for any comment on those findings.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 11/3/13 3:32 AM, Kiran Ayyagari a écrit :
On Sun, Nov 3, 2013 at 4:09 AM, Emmanuel Lécharny elecha...@gmail.comwrote:
Hi,
today, I investigated the way the Mavibot Partition is implemented, wrt
to the modifications I've made those past month. There are a few missing
parts, I'm afraid
,
Emmanuel Lécharny
www.iktek.com
though.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 11/10/13 10:25 AM, Emmanuel Lécharny a écrit :
Just quick heads up :
Im tryinh to fix a bug that forbid us to inject a few thousands entries
in the server (the injection test fails after aound 3000 thousands of
entries added). The problem lies in teh way we serialize/deserialize
Ok, pb fixed.
It was slightly more complex that I expected. The key length wasn't
stored for nodes, as shown in my previous mail, but the impact was a bit
more than just what I thought (ie, the fix was not what I suggested).
Le 11/11/13 8:14 AM, Emmanuel Lécharny a écrit :
Le 11/10/13 10:25 AM
, and 2,33 x faster
for searches... Will run a test with 100 000 entries.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
-
byte[]/byte[] - String conversions. The replace() method in index is
also extremelly important (we call drop/add as many times as we have
RDNs in an entry).
But the most important one would be the clone removal. It eats around
25% of the search time.
--
Regards,
Cordialement,
Emmanuel Lécharny
be to create special indexes in those special cases (ie, not always).
Most of the time, we will be good with standard index with hashed keys.
We just need to ask for those specific index to be created in the
configuration.
That is at least something to consider.
--
Regards,
Cordialement,
Emmanuel
error. One of the potential pb is that the
sub-btree we create for duplicate values aren't managed, and aren't read
back.
I'll investigate the pb tomorrow (it's a bit late now).
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
will always
point to teh first value of the found key.
That would make it easier for implementers to use te cursor lib...
Thoughs ?
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
there is a confusion between those two names. This is also true
for many classes :
Page, Leaf, Node, *Result, TupleCursorImpl, Transaction, ParentPos,
BulkDataSorter, BTreeFactory, BTreConfiguration, BTreeBuilder.
There is probably some renaming needed here.
--
Regards,
Cordialement,
Emmanuel Lécharny
,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Hi guys,
this morning, I launched a test on core-integ to see where we are, after
a few modifications I have made on Mavibot :
Tests run: 690, Failures: 6, Errors: 12, Skipped: 132
This is better than last week, but still there is some progress.
--
Regards,
Cordialement,
Emmanuel Lécharny
Le 11/24/13 10:51 AM, Emmanuel Lécharny a écrit :
Hi guys,
this morning, I launched a test on core-integ to see where we are, after
a few modifications I have made on Mavibot :
Tests run: 690, Failures: 6, Errors: 12, Skipped: 132
This is better than last week, but still there is some
Le 11/24/13 12:50 PM, Emmanuel Lécharny a écrit :
Le 11/24/13 10:51 AM, Emmanuel Lécharny a écrit :
Hi guys,
this morning, I launched a test on core-integ to see where we are, after
a few modifications I have made on Mavibot :
Tests run: 690, Failures: 6, Errors: 12, Skipped: 132
Le 11/28/13 7:26 PM, Emmanuel Lécharny a écrit :
Le 11/24/13 12:50 PM, Emmanuel Lécharny a écrit :
Le 11/24/13 10:51 AM, Emmanuel Lécharny a écrit :
Hi guys,
this morning, I launched a test on core-integ to see where we are, after
a few modifications I have made on Mavibot :
Tests run: 690
Le 11/29/13 10:31 AM, Emmanuel Lécharny a écrit :
Le 11/28/13 7:26 PM, Emmanuel Lécharny a écrit :
Le 11/24/13 12:50 PM, Emmanuel Lécharny a écrit :
Le 11/24/13 10:51 AM, Emmanuel Lécharny a écrit :
Hi guys,
this morning, I launched a test on core-integ to see where we are, after
a few
Le 11/30/13 1:25 AM, Emmanuel Lécharny a écrit :
Le 11/29/13 10:31 AM, Emmanuel Lécharny a écrit :
Le 11/28/13 7:26 PM, Emmanuel Lécharny a écrit :
Le 11/24/13 12:50 PM, Emmanuel Lécharny a écrit :
Le 11/24/13 10:51 AM, Emmanuel Lécharny a écrit :
Hi guys,
this morning, I launched a test
that could be done in a 2.0-RC2.
We are not far at all from a RC1 !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
: this is a test done with no network
involved).
Not bad...
Althouh the cache configuation now requires a lot of love in order to be
used for real, but this is not something big to do...
More to come after the next profiling session !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
part...
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 12/4/13 10:07 AM, Emmanuel Lécharny a écrit :
Hi guys,
I have added a warning on the linux download page for users
experimenting a crash, to avoid getting many reports (see
http://directory.staging.apache.org/studio/download/download-linux.html).
I guess that we would like to switch
Le 12/3/13 11:20 PM, Howard Chu a écrit :
Emmanuel Lécharny wrote:
Hi !
last numbers I got are quite interesting, now that we are corectly
leveraging the cache (alias cache, ParentIdAndRdn cache aka PIAR cache,
entry cache). Still, the way we configure and initialize the cache is
far from
^
|
+-- prev() : we are on 1
My point is that a next or a prev operation should move forward or
backward and return the current element *after* the move, not the opposite.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
,
Emmanuel Lécharny
www.iktek.com
Do not release Mavibot 1.0.0-M3
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
?
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 11/15/13 1:41 PM, Emmanuel Lécharny a écrit :
Le 11/15/13 12:08 PM, Kiran Ayyagari a écrit :
The name BTree is used for both in-memory as well as managed types
this might be slightly confusing unless one looks at the package names.
I suggest we name them as
1. BTree - the type of tree
Hi,
I close the vote, we have had 3 +1 :
Alex, Kiran and me.
This is quite a minimum...
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
, and when I profile the code, around 35% of the time is
spent in it). ANother important feature would be to use MemoryMapped
Files to increase the speed.
Ah, I almost forgot : documentation !!!
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
this problem, which would be a
gross exageration ;-)
This is really an exciting project with a lot of challenging problems ...
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
from the current one to the modified one by changing the first
page, so it's a one step operation.
Btw, the Node has been simplified : there is no need to keep the lentght
of page offsets, they are always 8 bytes long.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
the transaction feature is less
critical, even if it's good to have in the near future), and I *think*
we will be able to release ApacheDS with this backand as the main backend...
thoughts ?
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 12/20/13 11:09 PM, Howard Chu a écrit :
Emmanuel Lécharny wrote:
Hi guys,
transaction support is not easy, and while working on it, I realized
that the most critical part of it is the global update of the btrees
headers. This is a mandatory part of the transaction process, but it's
also
Le 12/20/13 9:21 AM, Emmanuel Lécharny a écrit :
After a second thought, it might be simpler to use 2 tables to store the
references to the current BTrees and to the modified btrees.
This second thought was palin wrong. We *need* a tree of btrees,
otherwise we won't be able to keep old
Le 12/27/13 5:28 PM, Marc Boorshtein a écrit :
If I start and stop ApacheDS cleanly then I have no issues. But if
ApacheDS is shutdown abruptly (ie the jvm crashes) I can't re-start it with
the following Exception:
Which version of ApacheDS ?
--
Regards,
Cordialement,
Emmanuel Lécharny
Le 12/27/13 5:58 PM, Marc Boorshtein a écrit :
On Fri, Dec 27, 2013 at 11:57 AM, Emmanuel Lécharny
elecha...@gmail.comwrote:
Le 12/27/13 5:28 PM, Marc Boorshtein a écrit :
If I start and stop ApacheDS cleanly then I have no issues. But if
ApacheDS is shutdown abruptly (ie the jvm crashes
can insolate some problem, we would be happy to check what's
going on !
One last thing : we will conduct some load tests very soon.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
for
the next 2 days (and my brain might be a bit too fried in the 2 or 3
days after those two days to be useful at all...)
Have a happy new year in the mean time !
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Hi guys, and an Happy New Year !
I'm currently working on revamping Mavibot to make it possible to use
revisions (something which is currently not available).
I have created a JIRA for that
(https://issues.apache.org/jira/browse/MAVIBOT-34)
The way it will be implemented implies a huge
also useless, as we already store
the number of children and descendants in the ParentIdAndRdn index.
I'm not 100% sure, because I never spent too much time in the logic of
this part of the server, so is anyone can confirm that I'm wrog ro right ?
Many thanks !
--
Regards,
Cordialement,
Emmanuel
Hi guys,
sorry for having been a bit silent those last 2 weeks... Xmas, NYD and
some personnal issues to deal with make it a bit difficult to move
forward on mavibot at the pace I would have liked to.
Here is a current status:
o I have refactored the code heavily so that we can now use
Hi guys,
I'm currently cleaning the code after my refactoring. Here are a
suggestion for a better naming of the elements we are using in Mavibot :
o B-tree should be the generic name in comments (instead of
BTree/Btree/btree/b-tree...)
Thoughts ?
--
Regards,
Cordialement,
Emmanuel Lécharny
Hi guys,
I have added the release of the Btree of Btrees copied pages, when we
commit a change. This is one more step forward.
The next step will be to move the copied pages into the CopiedPages btree.
Le 1/11/14 6:31 AM, Emmanuel Lécharny a écrit :
Hi guys,
sorry for having been a bit
...
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Next step : completed. I'm now able to get data from many revisions of a
B-tree !
Le 1/16/14 7:38 PM, Emmanuel Lécharny a écrit :
Done !
It was quite painful, with many small bugs that took a while to fix, due
to the lack of tooling atm (enough said I had to do many hexdump -C
mavibot.db
in the file.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
for the heads up, I will give it a try tonight
On Fri, Jan 17, 2014 at 11:19 PM, Emmanuel Lécharny
elecha...@gmail.comwrote:
Next step : completed. I'm now able to get data from many revisions of a
B-tree.
The insert and delete operations are now fully functional.
Most of the tests are now
Le 1/19/14 2:41 PM, Kiran Ayyagari a écrit :
Is anybody planning to attend ApacheCon (April 7-9, 2014) in Denver,
Colorado ?
Maybe. If Mavibot is advanced enough, I think it would be a good thing
to present it. Escimo would also worth some presentation.
--
Regards,
Cordialement,
Emmanuel
same for the update methods.
Le 1/21/14 1:44 AM, Emmanuel Lécharny a écrit :
I'm still fighting with the multi-threaded in-memory B-tree.
The way we handle the rootpage is not correct : we use it in search
operations while it can be modified in insert/delete operations. I tried
to protect
calm hours.
Le 1/22/14 11:53 PM, Emmanuel Lécharny a écrit :
I'm dumb... And this is plain normal, I currently don't have time to
focus on this part of the code, so I'm just trying to fix it the wrong
way (ie, doig this and that quite randomly).
We do have a ReadTransaction class which *must
,
Cordialement,
Emmanuel Lécharny
www.iktek.com
in OSGi, except that we don't have
time atm.
So my +1 to Kiran suggestion, and when we can spend some time to
understand the ins and outs of OSGi, we will rethink the full logic.
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
Le 2/7/14 11:22 AM, Kiran Ayyagari a écrit :
On Fri, Feb 7, 2014 at 3:42 PM, Emmanuel Lécharny elecha...@symas.comwrote:
Hi,
there are two things to consider in order to support transactions :
- we could have automatique transaction per operation (ie, we don't have
to start a transaction
701 - 800 of 2502 matches
Mail list logo
