Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2484
> @qihongxu @michaelandrepearce
> I'm running now a CI job: it will take some time, but when it will be
fine I will merge this
> @qihongxu After all the
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2484
@michaelandrepearce @franz1981
After we ran tests on both version (one with no lock and the other with new
LivePageCache & no lock ), the result chart is as below.
 |
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2484
> @michaelandrepearce @qihongxu Ok, checked: the latest version of this PR
+ my branch
https://github.com/franz1981/activemq-artemis/tree/lock-free-live-page-cache is
fully non-block
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2484
@michaelandrepearce
Removed readlock of isPaging().
Also as @franz1981 suggested now only volatile load
addressFullMessagePolicy once on each call.
Please review and notify
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2482
> Im very cautious of optimising for one case
> Its very unusual for consumers to go away as it is good design in most
mom a consumer is long lived. Like wise upgrade is a non
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2484
@michaelandrepearce
> @qihongxu you're using JMS api not core api then?
Yes (sometimes for for compatibility concern)
> @qihongxu i dont see checkDepage
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2482
@michaelandrepearce
There are some cases will perform lots of rollbacks in a short period of
time. For example if we would like to upgrade our server while thousands of
consumers
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2484
> Instead of transaction consumer you could use client acknowledge or even
individual acknowledge.
It seems that both client-acknowledge and individual-acknowledge m
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2484
> @qihongxu,
>
> re:
> "
> According to this we have an internal branch that removes isPaging()
methodâs readlock in PagingStoreImpl, along with a
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2484
@michaelandrepearce As presented in test result adding a specific executor
do improve producer rates, and also means page() method in PagingStoreImpl will
acquire writelock more often
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2482
> Nice catch!!!
> If you care about checking perf of paging please try to profile the
broker with https://github.com/jvm-profiling-tools/async-profiler using
lock/cpu even
GitHub user qihongxu opened a pull request:
https://github.com/apache/activemq-artemis/pull/2484
Use a specific executor for pageSyncTimer
Improve paging throughput by using a specific executor for pageSyncTimer
Improving throughput on paging mode is one of our concerns
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2482
@michaelandrepearce Thanks for the review. Now I have changed the code at
your suggestion.
---
GitHub user qihongxu opened a pull request:
https://github.com/apache/activemq-artemis/pull/2482
ARTEMIS-2214 Cache durable in PagedReference
We recently performed a test on artemis broker and found a severe
performance issue.
When paged messages are being consumed
Github user qihongxu commented on the issue:
https://github.com/apache/activemq-artemis/pull/2482
[stacks.txt](https://github.com/apache/activemq-artemis/files/2709657/stacks.txt)
---
Github user qihongxu commented on a diff in the pull request:
https://github.com/apache/activemq-artemis/pull/1894#discussion_r170405758
--- Diff:
artemis-server/src/main/java/org/apache/activemq/artemis/core/persistence/impl/journal/AbstractJournalStorageManager.java
GitHub user qihongxu opened a pull request:
https://github.com/apache/activemq-artemis/pull/1894
ARTEMIS-1700 Fixed deadlock in paging state
JournalStorageManager is not indeed using a `single` thread. We apply this
patch to use a simple single thread executor.
We have seen
17 matches
Mail list logo