You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cassandra.apache.org by Andrew Whang <an...@gmail.com> on 2017/07/28 04:37:13 UTC
CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
I'm wondering if anyone has been able to patch CASSANDRA-9472 to 3.0,
without breaking unit tests. The patch was introduced in 3.4, but 3.0.x
contains unit tests and code from later 3.x releases, which makes debugging
unit test failures difficult - i.e. SSTableCorruptionDetectionTest, which
was introduced in 3.7 and is found in 3.0.14, but not in 3.4.
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Andrew Whang <an...@gmail.com>.
Hi Jay,
Here's the backport to 3.0.14 -
https://github.com/whangsf/cassandra/commit/8db2e3ed412e42fed1da2d85ee7d086edcc8ae4c.
This should pass all unit tests, but please let me know if you have any
issues.
Thanks,
Andrew
On Mon, Aug 28, 2017 at 7:35 PM, Jay Zhuang <ja...@yahoo.com.invalid>
wrote:
> Hi Andrew,
>
> Do you mind sharing the backport patch? We're very interested in that,
> 20-30% improvement sounds great to us.
>
> Thanks,
> Jay
>
> On 7/27/17 11:52 PM, Andrew Whang wrote:
> > Yes, seeing latency improvement after backporting 9472 to 3.0.13. We are
> > measuring p99 latency, thus moving objects off heap improved gc stalls,
> > which directly affects our read/write p99 latency.
> >
> > On Thu, Jul 27, 2017 at 10:54 PM, Jeff Jirsa <jj...@gmail.com> wrote:
> >
> >> This is after you backported 9472 to 3.0?
> >>
> >> --
> >> Jeff Jirsa
> >>
> >>
> >>> On Jul 27, 2017, at 10:33 PM, Andrew Whang <an...@gmail.com>
> >> wrote:
> >>>
> >>> Jay,
> >>>
> >>> We see ~20% write latency improvement on 3.0.13 in a write-heavy
> >> workload,
> >>> using offheap_objects. offheap_buffers only offered minimal
> improvement.
> >>>
> >>> On Thu, Jul 27, 2017 at 10:06 PM, Jay Zhuang
> >> <ja...@yahoo.com.invalid>
> >>> wrote:
> >>>
> >>>> Hi Andrew,
> >>>>
> >>>> Do you see performance gain from reintroducing off-heap memtables for
> >>>> 3.0.x? When we were on 2.2.x we saw big improvements from enabling
> >>>> off-heap memtables.
> >>>>
> >>>> Thanks,
> >>>> Jay
> >>>>
> >>>>> On 7/27/17 9:37 PM, Andrew Whang wrote:
> >>>>> I'm wondering if anyone has been able to patch CASSANDRA-9472 to 3.0,
> >>>>> without breaking unit tests. The patch was introduced in 3.4, but
> 3.0.x
> >>>>> contains unit tests and code from later 3.x releases, which makes
> >>>> debugging
> >>>>> unit test failures difficult - i.e. SSTableCorruptionDetectionTest,
> >>>> which
> >>>>> was introduced in 3.7 and is found in 3.0.14, but not in 3.4.
> >>>>>
> >>>>
> >>>> ---------------------------------------------------------------------
> >>>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> >>>> For additional commands, e-mail: dev-help@cassandra.apache.org
> >>>>
> >>>>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> >> For additional commands, e-mail: dev-help@cassandra.apache.org
> >>
> >>
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> For additional commands, e-mail: dev-help@cassandra.apache.org
>
>
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Jay Zhuang <ja...@yahoo.com.INVALID>.
Hi Andrew,
Do you mind sharing the backport patch? We're very interested in that,
20-30% improvement sounds great to us.
Thanks,
Jay
On 7/27/17 11:52 PM, Andrew Whang wrote:
> Yes, seeing latency improvement after backporting 9472 to 3.0.13. We are
> measuring p99 latency, thus moving objects off heap improved gc stalls,
> which directly affects our read/write p99 latency.
>
> On Thu, Jul 27, 2017 at 10:54 PM, Jeff Jirsa <jj...@gmail.com> wrote:
>
>> This is after you backported 9472 to 3.0?
>>
>> --
>> Jeff Jirsa
>>
>>
>>> On Jul 27, 2017, at 10:33 PM, Andrew Whang <an...@gmail.com>
>> wrote:
>>>
>>> Jay,
>>>
>>> We see ~20% write latency improvement on 3.0.13 in a write-heavy
>> workload,
>>> using offheap_objects. offheap_buffers only offered minimal improvement.
>>>
>>> On Thu, Jul 27, 2017 at 10:06 PM, Jay Zhuang
>> <ja...@yahoo.com.invalid>
>>> wrote:
>>>
>>>> Hi Andrew,
>>>>
>>>> Do you see performance gain from reintroducing off-heap memtables for
>>>> 3.0.x? When we were on 2.2.x we saw big improvements from enabling
>>>> off-heap memtables.
>>>>
>>>> Thanks,
>>>> Jay
>>>>
>>>>> On 7/27/17 9:37 PM, Andrew Whang wrote:
>>>>> I'm wondering if anyone has been able to patch CASSANDRA-9472 to 3.0,
>>>>> without breaking unit tests. The patch was introduced in 3.4, but 3.0.x
>>>>> contains unit tests and code from later 3.x releases, which makes
>>>> debugging
>>>>> unit test failures difficult - i.e. SSTableCorruptionDetectionTest,
>>>> which
>>>>> was introduced in 3.7 and is found in 3.0.14, but not in 3.4.
>>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
>>>> For additional commands, e-mail: dev-help@cassandra.apache.org
>>>>
>>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
>> For additional commands, e-mail: dev-help@cassandra.apache.org
>>
>>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Jon Haddad <jo...@gmail.com>.
+1. IMO there’s very little reason to use 3.0 at this point. If someone wants to back port and make a 3.0 patch publicly available, cool, but merging it into 3.0 after 2 years doesn’t make much sense to me.
> On Jul 31, 2017, at 9:26 AM, Jeremiah D Jordan <je...@gmail.com> wrote:
>
>
>> On Jul 31, 2017, at 12:17 PM, Jeff Jirsa <jj...@apache.org> wrote:
>> On 2017-07-29 10:02 (-0700), Jay Zhuang <ja...@yahoo.com.INVALID> wrote:
>>> Should we consider back-porting it to 3.0 for the community? I think
>>> this is a performance regression instead of new feature. And we have the
>>> feature in 2.1, 2.2.
>>>
>>
>> Personally / individually, I'd much rather see 3.0 stabilize.
>
> +1. The feature is there in 3.11.x if you are running one of the use cases where this helps, and for most existing things 3.0 and 3.11 are about the same stability, so you can go to 3.11.x if you want to keep using the off heap stuff.
>
> -Jeremiah
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> For additional commands, e-mail: dev-help@cassandra.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Jeremiah D Jordan <je...@gmail.com>.
> On Jul 31, 2017, at 12:17 PM, Jeff Jirsa <jj...@apache.org> wrote:
> On 2017-07-29 10:02 (-0700), Jay Zhuang <ja...@yahoo.com.INVALID> wrote:
>> Should we consider back-porting it to 3.0 for the community? I think
>> this is a performance regression instead of new feature. And we have the
>> feature in 2.1, 2.2.
>>
>
> Personally / individually, I'd much rather see 3.0 stabilize.
+1. The feature is there in 3.11.x if you are running one of the use cases where this helps, and for most existing things 3.0 and 3.11 are about the same stability, so you can go to 3.11.x if you want to keep using the off heap stuff.
-Jeremiah
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Jeff Jirsa <jj...@apache.org>.
On 2017-07-29 10:02 (-0700), Jay Zhuang <ja...@yahoo.com.INVALID> wrote:
> Should we consider back-porting it to 3.0 for the community? I think
> this is a performance regression instead of new feature. And we have the
> feature in 2.1, 2.2.
>
Personally / individually, I'd much rather see 3.0 stabilize.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Jay Zhuang <ja...@yahoo.com.INVALID>.
Should we consider back-porting it to 3.0 for the community? I think
this is a performance regression instead of new feature. And we have the
feature in 2.1, 2.2.
On 7/27/17 11:52 PM, Andrew Whang wrote:
> Yes, seeing latency improvement after backporting 9472 to 3.0.13. We are
> measuring p99 latency, thus moving objects off heap improved gc stalls,
> which directly affects our read/write p99 latency.
>
> On Thu, Jul 27, 2017 at 10:54 PM, Jeff Jirsa <jj...@gmail.com> wrote:
>
>> This is after you backported 9472 to 3.0?
>>
>> --
>> Jeff Jirsa
>>
>>
>>> On Jul 27, 2017, at 10:33 PM, Andrew Whang <an...@gmail.com>
>> wrote:
>>>
>>> Jay,
>>>
>>> We see ~20% write latency improvement on 3.0.13 in a write-heavy
>> workload,
>>> using offheap_objects. offheap_buffers only offered minimal improvement.
>>>
>>> On Thu, Jul 27, 2017 at 10:06 PM, Jay Zhuang
>> <ja...@yahoo.com.invalid>
>>> wrote:
>>>
>>>> Hi Andrew,
>>>>
>>>> Do you see performance gain from reintroducing off-heap memtables for
>>>> 3.0.x? When we were on 2.2.x we saw big improvements from enabling
>>>> off-heap memtables.
>>>>
>>>> Thanks,
>>>> Jay
>>>>
>>>>> On 7/27/17 9:37 PM, Andrew Whang wrote:
>>>>> I'm wondering if anyone has been able to patch CASSANDRA-9472 to 3.0,
>>>>> without breaking unit tests. The patch was introduced in 3.4, but 3.0.x
>>>>> contains unit tests and code from later 3.x releases, which makes
>>>> debugging
>>>>> unit test failures difficult - i.e. SSTableCorruptionDetectionTest,
>>>> which
>>>>> was introduced in 3.7 and is found in 3.0.14, but not in 3.4.
>>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
>>>> For additional commands, e-mail: dev-help@cassandra.apache.org
>>>>
>>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
>> For additional commands, e-mail: dev-help@cassandra.apache.org
>>
>>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Andrew Whang <an...@gmail.com>.
Yes, seeing latency improvement after backporting 9472 to 3.0.13. We are
measuring p99 latency, thus moving objects off heap improved gc stalls,
which directly affects our read/write p99 latency.
On Thu, Jul 27, 2017 at 10:54 PM, Jeff Jirsa <jj...@gmail.com> wrote:
> This is after you backported 9472 to 3.0?
>
> --
> Jeff Jirsa
>
>
> > On Jul 27, 2017, at 10:33 PM, Andrew Whang <an...@gmail.com>
> wrote:
> >
> > Jay,
> >
> > We see ~20% write latency improvement on 3.0.13 in a write-heavy
> workload,
> > using offheap_objects. offheap_buffers only offered minimal improvement.
> >
> > On Thu, Jul 27, 2017 at 10:06 PM, Jay Zhuang
> <ja...@yahoo.com.invalid>
> > wrote:
> >
> >> Hi Andrew,
> >>
> >> Do you see performance gain from reintroducing off-heap memtables for
> >> 3.0.x? When we were on 2.2.x we saw big improvements from enabling
> >> off-heap memtables.
> >>
> >> Thanks,
> >> Jay
> >>
> >>> On 7/27/17 9:37 PM, Andrew Whang wrote:
> >>> I'm wondering if anyone has been able to patch CASSANDRA-9472 to 3.0,
> >>> without breaking unit tests. The patch was introduced in 3.4, but 3.0.x
> >>> contains unit tests and code from later 3.x releases, which makes
> >> debugging
> >>> unit test failures difficult - i.e. SSTableCorruptionDetectionTest,
> >> which
> >>> was introduced in 3.7 and is found in 3.0.14, but not in 3.4.
> >>>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> >> For additional commands, e-mail: dev-help@cassandra.apache.org
> >>
> >>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> For additional commands, e-mail: dev-help@cassandra.apache.org
>
>
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Jeff Jirsa <jj...@gmail.com>.
This is after you backported 9472 to 3.0?
--
Jeff Jirsa
> On Jul 27, 2017, at 10:33 PM, Andrew Whang <an...@gmail.com> wrote:
>
> Jay,
>
> We see ~20% write latency improvement on 3.0.13 in a write-heavy workload,
> using offheap_objects. offheap_buffers only offered minimal improvement.
>
> On Thu, Jul 27, 2017 at 10:06 PM, Jay Zhuang <ja...@yahoo.com.invalid>
> wrote:
>
>> Hi Andrew,
>>
>> Do you see performance gain from reintroducing off-heap memtables for
>> 3.0.x? When we were on 2.2.x we saw big improvements from enabling
>> off-heap memtables.
>>
>> Thanks,
>> Jay
>>
>>> On 7/27/17 9:37 PM, Andrew Whang wrote:
>>> I'm wondering if anyone has been able to patch CASSANDRA-9472 to 3.0,
>>> without breaking unit tests. The patch was introduced in 3.4, but 3.0.x
>>> contains unit tests and code from later 3.x releases, which makes
>> debugging
>>> unit test failures difficult - i.e. SSTableCorruptionDetectionTest,
>> which
>>> was introduced in 3.7 and is found in 3.0.14, but not in 3.4.
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
>> For additional commands, e-mail: dev-help@cassandra.apache.org
>>
>>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Andrew Whang <an...@gmail.com>.
Jay,
We see ~20% write latency improvement on 3.0.13 in a write-heavy workload,
using offheap_objects. offheap_buffers only offered minimal improvement.
On Thu, Jul 27, 2017 at 10:06 PM, Jay Zhuang <ja...@yahoo.com.invalid>
wrote:
> Hi Andrew,
>
> Do you see performance gain from reintroducing off-heap memtables for
> 3.0.x? When we were on 2.2.x we saw big improvements from enabling
> off-heap memtables.
>
> Thanks,
> Jay
>
> On 7/27/17 9:37 PM, Andrew Whang wrote:
> > I'm wondering if anyone has been able to patch CASSANDRA-9472 to 3.0,
> > without breaking unit tests. The patch was introduced in 3.4, but 3.0.x
> > contains unit tests and code from later 3.x releases, which makes
> debugging
> > unit test failures difficult - i.e. SSTableCorruptionDetectionTest,
> which
> > was introduced in 3.7 and is found in 3.0.14, but not in 3.4.
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> For additional commands, e-mail: dev-help@cassandra.apache.org
>
>
Re: CASSANDRA-9472 Reintroduce off heap memtables - patch to 3.0
Posted by Jay Zhuang <ja...@yahoo.com.INVALID>.
Hi Andrew,
Do you see performance gain from reintroducing off-heap memtables for
3.0.x? When we were on 2.2.x we saw big improvements from enabling
off-heap memtables.
Thanks,
Jay
On 7/27/17 9:37 PM, Andrew Whang wrote:
> I'm wondering if anyone has been able to patch CASSANDRA-9472 to 3.0,
> without breaking unit tests. The patch was introduced in 3.4, but 3.0.x
> contains unit tests and code from later 3.x releases, which makes debugging
> unit test failures difficult - i.e. SSTableCorruptionDetectionTest, which
> was introduced in 3.7 and is found in 3.0.14, but not in 3.4.
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
For additional commands, e-mail: dev-help@cassandra.apache.org