Github user michaelandrepearce closed the pull request at:
https://github.com/apache/activemq-artemis/pull/1262
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1263
ARTEMIS-1156: FIX: Long Autoboxing occurring on Hot Path
Building on ARTEMIS-905 JCtools ConcurrentMap replacement first proposed
but currently parked by @franz1981, replace
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
See Jira with profiler output showing long object creation caused by auto
boxing [https://issues.apache.org/jira/browse/ARTEMIS-1156](url)
---
If your project is set up for it
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@franz1981
On 1) compute isn't being used within this class, the methods used are
::get ::put ::contains ::containsKey, indeed this is why i state this is
tar
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@franz1981 just updated to add in the journal compactor.
I ran out of time today, to deploy and run this with a profiler on our
servers to verify that this hasn&#
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
No. Unfortunately not, there is an int version and a hashset for objects
(but that will cause autoboxing)
---
If your project is set up for it, you can reply to this email and
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
it is largely based on the NonBlockingHashSet that internally uses
NonBlockingHashMap taking object keys, but replaced the internal map with their
NonBlockingHashMapLong and
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
Yeah still got a few bits to clean up. Also I need to run it on server
under load I ran out of time today. But review comments would be good
---
If your project is set up for
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
Well there's another fix I have locally I need to commit to fix the
distribution as it's missing the dependency in the zip/tar go. But on the move
atm
---
If your
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
dep.xml updated and check style fixed.
have managed to get results back from testing, allocation of longs is
greatly reduced (oddly theres one autobox in
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
https://cloud.githubusercontent.com/assets/1387822/25913734/b9930f5e-35b3-11e7-84af-a4073e0643c3.png";>
https://cloud.githubusercontent.com/assets/1387822/25913736/
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
from 127mb prior and now just 11mb. :), and i think we can remove that also
:) might have a look what the Integer is upto, in the allocations , it looks
like it auto boxing
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
ok so the int, is due to connection.getID return Object, for the two
implementations netty it is int, and InVM its a String UUID. I would suggest
the getID becomes an int on
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
So I'm just doing a bit more of a high load, seems the hash set, as unlike
the maps used in the JournalImpl which are fairly permanent, the sets are
temporal they'
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
If you could hold of merging. I am wanting to see by changing the
collections to ones used in Apache bookkeeper if we get any better.
---
If your project is set up for it, you
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@clebertsucnic
Just seen what you did, oh I like the idea what you did instead of going to
primitive, simply go the other way an make it an object and avoid constant
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
this will only occur where the primitive:boxed value is cached, the jvm
will cache/inline upto only a certain amount. If you have a constantly
increasing number like in this
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
```
public class Boxing {
public static void main(String... args){
//ensure we max'd out the caching magic.
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
So we have:
1. primitive collections
1. id's as objects (Long in this case)
1. flywheel/interning caching pattern
For 1:
we have tried, looks
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
So I've just had dived through a few flows very quickly using tracing,
there is one key area if we change from long to Long and that is in the
protocol/buffer exch
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@franz1981 As noted above, i wanted to look at the collections that
bookkeeper uses, the benefit here with these seems to be much reduced outside
tlab allocation of long[] when
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
correct me if I'm wrong, but these are not concurrent.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
FYI:
https://issues.apache.org/jira/browse/BOOKKEEPER-964
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
retest this please.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@franz1981 @clebertsuconic
So all updated to use the collections from bookkeeper. ran several load
tests through with and without changes, defiantly reduced the Long object
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
But I've removed deps on jctools entirely. Let me check your branch.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitH
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
So i see your change :), but i don't think the build issue isn't that.
As noted i have removed the dependency on jctools entirely, as ended up
using the
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
error is:
```
[ERROR] [ERROR] Some problems were encountered while processing the POMs:
[ERROR] Unresolveable build extension: Plugin
org.apache.felix:maven
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
No worries at all, if i have any divine inspiration i will let you know, if
not thanks in advance ð
---
If your project is set up for it, you can reply to this email and
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
Seems build has got past that issue. Doing nothing at all. Weird. Anyhow
now it's building again looks like I accidentally added some white space in a
stress test class
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@clebertsuconic yeah, checkstyle came back on the build, earlier just
wasn't even doing anything and just consistently complained
```
ERROR] [ERROR] Some pro
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
all green now :)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@clebertsuconic if you and @franz1981 want to have a final go over, once
happy if you need me to let me know ill rebase and squash for you so it can be
merged.
---
If your
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
Thanks :), I'm out for the evening tonight. But will look into firsts thing
in the morning.
Sent from my iPhone
> On 11 May 2017, at 21:49, cleber
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@clebertsuconic, have patched in the hashset fix, thanks for this. I also
made the same change on the hashmap. I have squashed and rebased.
I agree the move of the
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@clebertsuconic Don't worry figured it out :) managed to commit the moving
collections to the new collections package as you.
---
If your project is set up for it, yo
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
fyi, doing initial tests looks like still all good to me on the auto boxing
of long front.
---
If your project is set up for it, you can reply to this email and have your
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1263
@clebertsuconic great news. And def no worries on doing tweeks for git
history, your git fu far beats mine :)
thanks to you and also @franz1981 for the feedback
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1268
ARTEMIS-1162: Make new TimedBuffer Configurable
Make new Adapting TimedBuffer and old Fixed TimedBuffer configurable.
Rename new version of TimedBuffer to
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1268
@franz1981 I hope you don't mind me doing this, just as the new timed
buffer work you've done, whilst benchmarked we won't get any proper real use
feedback t
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1268
I actually started with a boolean flag option, but then thought about the
other bits being added and before we know it we'd have quite a few flags which
equate to lo
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1268
"paeans" wow my iphone is doing its best today!
I meant parameters
---
If your project is set up for it, you can reply to this email and have your
reply
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1268
I agree with you, thanks for the comments :) it's always good to have the
second opinions :). Let me re work it and I'll msg once ready for another
review :)
-
Github user michaelandrepearce closed the pull request at:
https://github.com/apache/activemq-artemis/pull/1268
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1268
@franz1981 I did re work this, but based on the discussion thread i may
kill it and the jira as clement is for an all or nothing approach, but with the
caveat of having a lot
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1268
@franz1981 and here in lies the issue, that the project doesn't currently
have a set of agreed performance metrics for different end 2 end scenarios,
that are run automati
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1268
I like the first link, much clearer and concisely written to explain the
idea.
Just need to prove and get the stats like in the second :)
---
If your project is set
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1270
ARTEMIS-1164: NameNotFoundException java.naming.provider.url to set url via
jndi
Support setting PROVIDER_URL on initial context to create default
connection factories. To
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1268
FYI I've closed the JIRA as won't fix.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your projec
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1270
@jbertram I was actually changing quite a bit still :( can you hold off
merge?
---
If your project is set up for it, you can reply to this email and have your
reply appear on
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1270
i see you merged, ill open a new pr, with the changes
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1271
ARTEMIS-1164: NameNotFoundException java.naming.provider.url to set url via
jndi
Support setting PROVIDER_URL on initial context to create default
connection factories
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1271
@jbertram could you merge this one if happy please, sorry about that had my
head in code this evening adding the extra test cases after you pointed me to
the existing ones
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1272
ARTEMIS-1129: Client Dependencies
Create shaded versions of the clients, so that end users have a single
clean dependency to depend on.
Third party dependency'
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1271
rebased, after https://github.com/apache/activemq-artemis/pull/1269 merged.
---
If your project is set up for it, you can reply to this email and have your
reply appear on
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1272
Re the issue.
We are able to run against remote broker fine.
are you running a remote client or invm client?
The invm parts aren't shaded as if y
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1272
I think I've managed to re-create your issue, i have updated the commit,
with the additional change in the all pom, that fixes this. (it is as you
suspect a repackage
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1272
@gemmellr as a project it is i think good citizenship to be able to provide
a clean artefact.
This is no different to netty, which guess what we use for the benefit of
this
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1271
@jbertram nudge :)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1272
Ive updated all the standard examples, that just use the client, to use the
new shaded single jar client.
---
If your project is set up for it, you can reply to this email and
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1272
@gemmellr actually there are some third party deps shaded in the netty one
also, e.g. jctools.
---
If your project is set up for it, you can reply to this email and have your
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1271
@jbertram thanks :)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1272
@clebertsuconic @jbertram with now decent test case range, using the
examples, and all looks to be passing still on PR build.
Anything else to do? (@clebertsuconic, i
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1278
@clebertsuconic def agree that it should be git ignore , as it is generated
at build time.
are you saying by adding this exclusion in rat list, the tests break, or
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1278
@clebertsuconic I'm happy with this change.
I have a few more tidy up bits, been working on, as have been exploding the
jars to check whats copied in and some bits
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1282
ARTEMIS-1129: Fixing tests after client all
To add ontop of the reduced-dependency.pom fix.
Clean up shaded jar
Ensure aggregated notice
Ensure correct license
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1282
@clebertsuconic re-pointed the PR to master.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1278
@clebertsuconic just seen this, rebased and re-raised against master
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1286
Can there be added a test case that this is fixing.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1285
Can a test case be added to ensure this doesn't regress in future?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitH
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1287
ARTEMIS-1179: Add Optional Client JMS Destination Cache
Add topic and queue cache maps in Session.
Add configuration to use cache or not with defaulting to false, which
Github user michaelandrepearce closed the pull request at:
https://github.com/apache/activemq-artemis/pull/1287
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1288
ARTEMIS-1179: Add Optional Client JMS Destination Cache
Add topic and queue cache maps in Session.
Add configuration to use cache or not with defaulting to false, which
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
build is yellow, and says success , not sure why its marked failed.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1290
ARTEMIS-1182: Add / Port in support for JNDIReferenceFactory
and JNDIStorable
* Add clones/copies of these class's from Qpid (though identical to
ActiveMQ 5.x
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
When a session is closed it's thrown away right, and you make a new
session. Would that not be redundant?
---
If your project is set up for it, you can reply to this emai
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
As in we don't hold any remote resource (which typically close methods are
for, so if no longer referenced GC would clear it out
---
If your project is set up for it, yo
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
@jbertram fair enough, just added clearing of them at the end of close().
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1290
again build output says success, jenkins icon is orange, meaning success of
this build.
---
If your project is set up for it, you can reply to this email and have your
reply
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
build output is success, and build is orange as in this build was
successful.
---
If your project is set up for it, you can reply to this email and have your
reply appear on
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
oh yeay sevntu not available again
[ERROR] Unresolveable build extension: Plugin
org.apache.felix:maven-bundle-plugin:3.3.0 or one of its dependencies could not
be
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
just pushed a commit comment change to make it rebuild to see if sevntu
wants to sort itself.
---
If your project is set up for it, you can reply to this email and have your
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
see it now you point it out, annoying as build shows end output as success
when watching console
https://builds.apache.org/job/ActiveMQ-Artemis-PR-Build/3081/console
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1290
did a rebuild by force pushing comment change, all seems green.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
For RA the property has use, as it doesn't have the same behaviour as the
connection factory so updated the test to reflect that.
---
If your project is set up for it
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
actually hold that thought, whilst i don't expect there to be the same
issue (aka Spring etc doing stuff), i can add it, so will add the support
incase some framework doe
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
@clebertsuconic all added and all looking good now with build :) thanks for
the tip.
---
If your project is set up for it, you can reply to this email and have your
reply
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
Indeed it's quite easy to get an account to publish on to central just a
simple jira ticket, so they should be willing to publish there.
If they're don'
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1288
cool, thanks for pushing it.
fyi, i think we found your presentation on youtube :)
https://www.youtube.com/watch?v=qF1Op1ZNBAk
---
If your project is set up for it
Github user michaelandrepearce commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/1295#discussion_r118983050
--- Diff:
artemis-cli/src/main/java/org/apache/activemq/artemis/cli/commands/tools/journal/StatJournal.java
---
@@ -0,0 +1,288
Github user michaelandrepearce commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/1295#discussion_r118982923
--- Diff:
artemis-journal/src/main/java/org/apache/activemq/artemis/core/io/buffer/counters/Profiler.java
---
@@ -0,0 +1,72
Github user michaelandrepearce commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/1295#discussion_r118982819
--- Diff:
artemis-journal/src/main/java/org/apache/activemq/artemis/core/io/buffer/TimedBuffer.java
---
@@ -93,10 +82,10
Github user michaelandrepearce commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/1295#discussion_r118983292
--- Diff:
artemis-journal/src/main/java/org/apache/activemq/artemis/core/io/buffer/TimedBuffer.java
---
@@ -93,10 +82,10
Github user michaelandrepearce commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/1295#discussion_r118982620
--- Diff:
artemis-journal/src/main/java/org/apache/activemq/artemis/core/io/buffer/TimedBuffer.java
---
@@ -288,7 +266,14 @@ private
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1295
Looks good +1, comments are just comments, I'm happy with as is.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as
GitHub user michaelandrepearce opened a pull request:
https://github.com/apache/activemq-artemis/pull/1299
ARTEMIS-1188: Update sevntu to 1.24.0 available in maven central
remove custom repo
update groupid to match artifact in maven central.
bump version also to that now
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1299
FYI - I'm happy to sort prolific check style violations now picked up on
upgrade of checkstyle to latest (key ones are redundant final modifiers and
also do while
Github user michaelandrepearce commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/1295#discussion_r118992513
--- Diff:
artemis-journal/src/main/java/org/apache/activemq/artemis/core/io/buffer/counters/Profiler.java
---
@@ -0,0 +1,72
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1295
PR raised to hopefully end the sevntu curse (aka use new version published
now in maven central)
https://github.com/apache/activemq-artemis/pull/1299
---
If your project
Github user michaelandrepearce commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/1295#discussion_r119039588
--- Diff:
artemis-journal/src/main/java/org/apache/activemq/artemis/core/io/buffer/TimedBuffer.java
---
@@ -93,10 +82,10
601 - 700 of 1769 matches
Mail list logo
