Mailing-List: contact user-help@kudu.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@kudu.apache.org
MIME-Version: 1.0
In-Reply-To: <CADY20s41OUGp=2wdNmyHwQ4U_ZKxDggm2OjqP6tXewDFHat29Q@mail.gmail.com>
References: <CAO3rhrTOmJsczX1gM9ZjrRa2iuvVRYkN8PeYgj5pw2L+NtjPTQ@mail.gmail.com>
 <CALo2W-WyYK1cRga9ooirL=uVsRbF_Y27hoK2wYcOu_h7dbkDLw@mail.gmail.com> <CADY20s41OUGp=2wdNmyHwQ4U_ZKxDggm2OjqP6tXewDFHat29Q@mail.gmail.com>
From: Jason Heo <jason.heo.sde@gmail.com>
Date: Mon, 10 Apr 2017 10:38:32 +0900
Message-ID: <CAO3rhrTp9ZJ5-GC41oTN2_RJsuoH5me9u2_+aK5FMeLAtd1r0Q@mail.gmail.com>
Subject: Re: How to flush `block_cache_capacity_mb` easily?
To: user@kudu.apache.org
Content-Type: multipart/alternative; boundary=94eb2c0b61e878651f054cc608ee
archived-at: Mon, 10 Apr 2017 01:38:37 -0000

--94eb2c0b61e878651f054cc608ee
Content-Type: text/plain; charset=UTF-8

Hi Todd.

I hope you had a good weekend.

Exactly, I'm testing the latency of cold-cache reads from SATA disks and
performance of difference schema designs as well.

We currently using Elasticsearch for a analytic service. ES has a "clear
cache API" feature, it makes me easy to test.

Thanks.

Jason.

2017-04-08 5:05 GMT+09:00 Todd Lipcon <todd@cloudera.com>:

> Hey Jason,
>
> Can I ask what the purposes of the testing is?
>
> One thing to note is that we're currently leaving a fair bit of
> performance on the table for cold-cache reads from spinning disks. So, if
> you find that the performance is not satisfactory, it's worth being aware
> that we will likely make some significant improvements in this area in the
> future.
>
> https://issues.apache.org/jira/browse/KUDU-1289 has some details.
>
> -Todd
>
> On Fri, Apr 7, 2017 at 8:44 AM, Dan Burkert <danburkert@apache.org> wrote:
>
>> Hi Jason,
>>
>> There is no command to have Kudu evict its block cache, but restarting
>> the tablet server process will have that effect.  Ideally all written data
>> will be flushed before the restart, otherwise startup/bootstrap will take a
>> bit longer. Flushing typically happens within 60s of the last write.
>> Waiting for flush and compaction is also a best-practice for read-only
>> benchmarks.  I'm not sure if someone else on the list has an easier way of
>> determining when a flush happens, but I typically look at the 'MemRowSet'
>> memory usage for the tablet on the /mem-trackers HTTP endpoint; it should
>> show something minimal like 256B if it's fully flushed and empty.  You can
>> also see details about how much memory is in the block cache on that page,
>> if that interests you.
>>
>> - Dan
>>
>> On Thu, Apr 6, 2017 at 11:23 PM, Jason Heo <jason.heo.sde@gmail.com>
>> wrote:
>>
>>> Hi.
>>>
>>> I'm using Apache Kudu 1.2 on CDH 5.10.
>>>
>>> Currently, I'm doing a performance test of Kudu.
>>>
>>> Flushing OS Page Cache is easy, but I don't know how to flush
>>> `block_cache_capacity_mb` easily.
>>>
>>> I currently execute SELECT statement over a unnecessarily table to evict
>>> cached block of testing table.
>>>
>>> It is cumbersome, so I'd like to know is there a command for flushing
>>> block caches (or another kudu's caches which I don't know yet)
>>>
>>> Thanks.
>>>
>>> Regards,
>>> Jason
>>>
>>
>>
>
>
> --
> Todd Lipcon
> Software Engineer, Cloudera
>

--94eb2c0b61e878651f054cc608ee
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi Todd.<div><br></div><div>I hope you had a good weekend.=
</div><div><br></div><div>Exactly, I&#39;m testing the latency of cold-cach=
e reads from SATA disks and performance of difference schema designs as wel=
l.</div><div><br></div><div>We currently using Elasticsearch for a analytic=
 service. ES has a &quot;clear cache API&quot; feature, it makes me easy to=
 test.</div><div><br></div><div>Thanks.</div><div><br></div><div>Jason.</di=
v></div><div class=3D"gmail_extra"><br><div class=3D"gmail_quote">2017-04-0=
8 5:05 GMT+09:00 Todd Lipcon <span dir=3D"ltr">&lt;<a href=3D"mailto:todd@c=
loudera.com" target=3D"_blank">todd@cloudera.com</a>&gt;</span>:<br><blockq=
uote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc =
solid;padding-left:1ex"><div dir=3D"ltr">Hey Jason,<div><br></div><div>Can =
I ask what the purposes of the testing is?</div><div><br></div><div>One thi=
ng to note is that we&#39;re currently leaving a fair bit of performance on=
 the table for cold-cache reads from spinning disks. So, if you find that t=
he performance is not satisfactory, it&#39;s worth being aware that we will=
 likely make some significant improvements in this area in the future.</div=
><div><br></div><div><a href=3D"https://issues.apache.org/jira/browse/KUDU-=
1289" target=3D"_blank">https://issues.apache.org/<wbr>jira/browse/KUDU-128=
9</a> has some details.</div><div><br></div><div>-Todd</div></div><div clas=
s=3D"gmail_extra"><div><div class=3D"h5"><br><div class=3D"gmail_quote">On =
Fri, Apr 7, 2017 at 8:44 AM, Dan Burkert <span dir=3D"ltr">&lt;<a href=3D"m=
ailto:danburkert@apache.org" target=3D"_blank">danburkert@apache.org</a>&gt=
;</span> wrote:<br><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 =
.8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir=3D"ltr">Hi Jason=
,<div><br></div><div>There is no command to have Kudu evict its block cache=
, but restarting the tablet server process will have that effect.=C2=A0 Ide=
ally all written data will be flushed before the restart, otherwise startup=
/bootstrap will take a bit longer. Flushing typically happens within 60s of=
 the last write.=C2=A0 Waiting for flush and compaction is also a best-prac=
tice for read-only benchmarks.=C2=A0 I&#39;m not sure if someone else on th=
e list has an easier way of determining when a flush happens, but I typical=
ly look at the &#39;MemRowSet&#39; memory usage for the tablet on the /mem-=
trackers HTTP endpoint; it should show something minimal like 256B if it=
9;s fully flushed and empty.=C2=A0 You can also see details about how much =
memory is in the block cache on that page, if that interests you.</div><spa=
n class=3D"m_5574670644068612997HOEnZb"><font color=3D"#888888"><div><br></=
div><div>- Dan</div></font></span></div><div class=3D"m_5574670644068612997=
HOEnZb"><div class=3D"m_5574670644068612997h5"><div class=3D"gmail_extra"><=
br><div class=3D"gmail_quote">On Thu, Apr 6, 2017 at 11:23 PM, Jason Heo <s=
pan dir=3D"ltr">&lt;<a href=3D"mailto:jason.heo.sde@gmail.com" target=3D"_b=
lank">jason.heo.sde@gmail.com</a>&gt;</span> wrote:<br><blockquote class=3D=
"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding=
-left:1ex"><div dir=3D"ltr">Hi.<div><br></div><div>I&#39;m using Apache Kud=
u 1.2 on CDH 5.10.</div><div><br></div><div>Currently, I&#39;m doing a perf=
ormance test of Kudu.</div><div><br></div><div>Flushing OS Page Cache is ea=
sy, but I don&#39;t know how to flush `block_cache_capacity_mb` easily.</di=
v><div><br></div><div>I currently execute SELECT statement over a unnecessa=
rily table to evict cached block of testing table.</div><div><br></div><div=
>It is cumbersome, so I&#39;d like to know is there a command for flushing =
block caches (or another kudu&#39;s caches which I don&#39;t know yet)</div=
><div><br></div><div>Thanks.</div><div><br></div><div>Regards,</div><div>Ja=
son</div></div>
</blockquote></div><br></div>
</div></div></blockquote></div><br><br clear=3D"all"><div><br></div></div><=
/div><span class=3D"HOEnZb"><font color=3D"#888888">-- <br><div class=3D"m_=
5574670644068612997gmail_signature" data-smartmail=3D"gmail_signature">Todd=
 Lipcon<br>Software Engineer, Cloudera</div>
</font></span></div>
</blockquote></div><br></div>

--94eb2c0b61e878651f054cc608ee--