Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
Many thanks Clebert.
Had a brief look over tonight. Seems good. Will spend a bit more time
tomorrow and also take the chance to update my fork.
Sent from my iP
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
Please review 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 does not have this feature
ena
Not finding epoll pakcets on netty connector.
Now that you mentioned. I think rebuilt netty. And I provably did
something wrong.
Will redo in the morning with a fresh Maven. Going to sleep now :)
On Wed, Mar 22, 2017 at 11:27 PM Michael André Pearce <
michael.andre.pea...@me.com> wrote:
>
That's odd re Mac not building for you, my laptop is MacBook and primary dev
env , as noted i rebuilt to recheck using a fresh clone and wiped my local
maven repo, I also built on fedora using aws.
What is the extract failure you're seeing?
> On 23 Mar 2017, at 02:58, Clebert Suconic wrote
I can't make it to build on mac and windows.
I need to make some changes also as we talked. I was using this to test
the proton changes. It works and builds on Linux. Need to fix the build on
mac and windows now.
On Wed, Mar 22, 2017 at 10:38 PM Michael André Pearce <
michael.andre.pea...@m
Hi Clebert
Any luck getting it building?
Did you take my branch or did you take the pr and apply to a local fork/branch
with other changes? If so can you point me to that so I can see what is up.
Also re the amqp stuff, I see a lot of upstream changes if I take latest now is
that expected to
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@clebertsuconic i just did a complete fresh clone and rebuild of this PR/My
branch which i rebased the other day. It still built ok for me. Should i rebase
again? Maybe some com
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
No haven't touched it
---
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 clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce did you ammend anything on your PR? I can't build it
any longer (especially outside of Linux.. like on a mac).
---
If your project is set up for it, you can rep
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
I think I got to the bottom of this... there was a leak.. and some proton
miss use.
---
If your project is set up for it, you can reply to this email and have your
reply appear on
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce that's what I was going after.. some sort of leak while
dealing with AMQP.
I will review it this week.
---
If your project is set up for it, you can re
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
Also:
#1:
io.netty.buffer.AdvancedLeakAwareByteBuf.writeBytes(AdvancedLeakAwareByteBuf.java:600)
io.netty.buffer.AbstractByteBuf.readBytes(AbstractByteBu
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
i turned on the netty leak detector:
22:55:59,492 SEVERE [io.netty.util.ResourceLeakDetector] LEAK:
ByteBuf.release() was not called before it's garbage-collected. See
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@clebertsuconic i have re-run with march larger run size, and now able to
reproduce.
So it seems there is a direct memory leak, this occurs with both epoll and
nio set
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce you have to follow all the dependencies on quiver.
probably the cpp didn't finish compilation on make?
There's something going on.. I will review it
Github user franz1981 commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce thanks for the results!
---
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
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
the only comment is with epoll, i notice the producer was producing faster
than the consumer during the test as such some queue depth occured, with nio
producer and consumer spe
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
Hi @clebertsuconic i managed to get it running.
Some notes:
I was not able to run your command as is, using the version that comes via
dnf / repo, seems some options
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce I couldn't check other options.. I was using AMQP..
so, I was using this command line on quiver to replicate it:
```
quiver q0 --sender qpid
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@clebertsuconic thanks, ill try look into also see if i spot anything.
I assume this is repeatable.
Was using the amqp client, looking at the quiver dependencies not
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce I did some performance tests with
https://github.com/ssorj/quiver and the test hung... I will have to take double
care before merging this.. it may take some ext
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce 2.0.0 (Artemis) is already cut. 3 days ago.. it was
just the vote on...
So this will be for 2.0.1.
I was thinking about setting the default
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@clebertsuconic I was thinking this afternoon has 2.0.0 been cut for
tagging? If not then maybe worth trying to get the property change done before
that as a 2.0.0 is a major br
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce nope.. I just procrastinated for an afternoon :) give
me till tomorrow
---
If your project is set up for it, you can reply to this email and have your
reply app
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@clebertsuconic I notice this didn't get merged still. Did I miss
understand and you expected me to make the configuration changes? if so no
worries, just let me know.
---
If
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@clebertsuconic sounds good.
---
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
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce I will add a new generic parameter, and deprecate the
old one.
Will do that on a separate commit. I just wanted to know if you had any
reason.
---
If
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@clebertsuconic I though about this but annoyingly theyre named nioBlah as
such just using that would be not directly indicate you're affecting epoll.
Like wise if we were to st
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@michaelandrepearce I don't think you really need three parameters for
this... just one would do...
- useEpoll...
the other two you introduced could just use t
Github user michaelandrepearce commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
@clebertsuconic I have just now rebased and squashed, hope this is what you
were wanting me to do.
---
If your project is set up for it, you can reply to this email and have yo
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
Awesome work! Testsuite is good...
---
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
Github user clebertsuconic commented on the issue:
https://github.com/apache/activemq-artemis/pull/1093
I am running the testsuite, and if everything is ok I should merge it
tomorrow...
I have 50% of tests running and so far so good.. will know in 2 hours...
32 matches
Mail list logo