Guys, I replied in ticket. Overall I liked the changes but I think this
needs additional elaboration. Please see
https://issues.apache.org/jira/browse/IGNITE-4828
--Yakov
2017-03-31 20:09 GMT+03:00 Denis Magda :
> Michael, thanks.
>
> I did minor improvements and merged them to IGNITE-4828 branc
Michael, thanks.
I did minor improvements and merged them to IGNITE-4828 branch and triggered
TeamCity tests:
http://ci.ignite.apache.org/viewQueued.html?itemId=526101&tab=queuedBuildOverviewTab
*Michael*, please check the tests results lately. I won’t be available in the
nearest 4 days.
*Ilya
Vova,
Not sure I understand what you mean. I believe that we just agreed that if
the partition count is set to the power of 2, then we can improve the
performance with a better hashing algorithm.
D.
On Fri, Mar 31, 2017 at 1:46 AM, Vladimir Ozerov
wrote:
> In order to overcome this problem, Ha
In order to overcome this problem, Hazelcast guys set non-power-of-two
partition count by default - 257.
On Fri, Mar 31, 2017 at 9:59 AM, michael.griggs wrote:
> The change is now ready for review:
>
> https://github.com/apache/ignite/pull/1707
>
>
>
> --
> View this message in context: http://a
The change is now ready for review:
https://github.com/apache/ignite/pull/1707
--
View this message in context:
http://apache-ignite-developers.2346864.n4.nabble.com/Distribution-of-keys-to-partitions-tp15455p16005.html
Sent from the Apache Ignite Developers mailing list archive at Nabble.com.
I created a PR to implement this. I ran the TC tests, but there are a lot of
errors. However, the errors seem unrelated to the change.
I see that other PRs are suffering with similar test failures. Have some
tests been broken by new 2.0 functionality and not fixed yet?
http://ci.ignite.apach
Excellent discovery, thanks Michael!
I would suggest doing the following. If we see that a number of partitions is a
power of two then the new algorithm will be applied, otherwise the warning will
be printed out and the *old* one approach will be used. Does this resolver all
the concerns?
Mich
On Wed, Mar 15, 2017 at 9:51 AM, Michael Griggs wrote:
> Have we ever heard of somebody needing to set the partition count to a
> non-power-of-two number? Perhaps we could restrict the method so that it
> will only accept a power of two as the partition count?
>
As Valentin suggested, we should
]
Sent: 15 March 2017 16:22
To: dev@ignite.apache.org
Subject: Re: Distribution of keys to partitions
Andrey,
Absolutely, your point is correct. I'm talking about default behavior which
must be as effective as possible. In case we do this optimization, I would also
show a warning if numb
Andrey,
Absolutely, your point is correct. I'm talking about default behavior which
must be as effective as possible. In case we do this optimization, I would
also show a warning if number of partitions is not a power of two.
-Val
On Wed, Mar 15, 2017 at 5:09 PM, Andrey Gura wrote:
> Anyway, w
Valentin Kulichenko wrote
> In 99% of cases number of partition is a power of two, because it's the
> default value. Almost no one changes it. If this change actually provides
> better distribution, it absolutely makes sense to do it.
>
> Michael, can you create a Jira ticket and put you findings
Anyway, we can't always use this optimization because it will not work
for non power of two values.
On Wed, Mar 15, 2017 at 6:48 PM, Valentin Kulichenko
wrote:
> In 99% of cases number of partition is a power of two, because it's the
> default value. Almost no one changes it. If this change actua
In 99% of cases number of partition is a power of two, because it's the
default value. Almost no one changes it. If this change actually provides
better distribution, it absolutely makes sense to do it.
Michael, can you create a Jira ticket and put you findings there?
-Val
On Wed, Mar 15, 2017 a
Michael,
it makes sense only for cases when partitions count is power of two.
Affinity function doesn't have this limitation.
Bu, of course, we can check, that partitions count is power of two and
use optimized hash code calculation.
On Wed, Mar 15, 2017 at 4:09 PM, Michael Griggs
wrote:
> Hi
14 matches
Mail list logo
