Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
MIME-Version: 1.0
In-Reply-To: 
 <CACCLA55ti9ey8PdDXG5Cn8G5=sdH3UdGTdLS9ZMQdU1gi2iEKQ@mail.gmail.com>
References: 
 <CACCLA57rnL5OyWYFjW_e+RGpA8Dt8DKBhd2YvF3Biy2MT3MCnw@mail.gmail.com>
 <CAEDUwd2XyQ+_B41hKhx95b47ozngv-4=vti3mrdHL598XXN-0w@mail.gmail.com>
 <CACCLA55ti9ey8PdDXG5Cn8G5=sdH3UdGTdLS9ZMQdU1gi2iEKQ@mail.gmail.com>
From: Alain RODRIGUEZ <arodrime@gmail.com>
Date: Fri, 21 Aug 2015 11:37:41 +0200
Message-ID: 
 <CA+VSrLoEY16zVG84RbRPdU0_RmxEDruvgr_039yTfsstZqsviQ@mail.gmail.com>
Subject: Re: GossipingPropertyFileSnitch vs Ec2Snitch
To: user@cassandra.apache.org
Content-Type: multipart/alternative; boundary=001a1133a02e10c26a051dcf06f6

--001a1133a02e10c26a051dcf06f6
Content-Type: text/plain; charset=UTF-8

Hi guys,

I am also wondering about this kind of stuff, since we are going out of AWS
and were using EC2Snitch. I felt free to write it here, since it looks
linked to previous comments.

@John, we have been using EC2Snitch for many years, and it just works fine
! Plus, we don't have to keep the GPFS file up to date every time we add a
node or remove one. EC2Snitch is quite strait-forward and easy to use. The
first "issue" we have today, is indeed to add a new DC from outside AWS. I
am trying to find out how to do that right now. If we figure out that this
is easy, you can perfectly go with EC2Snitch imho.

So what would be the procedure to go from EC2Snitch to GPFS ?

I see 2 possibilities and I am not sure of the result of each of these:

- Node by node, on a rolling restart way, and making sure that you use the
former DC names and racks. Copying the EC2 network. Would this be doable or
would that need a full cluster restart at once ?
- Add the new DC using directly GPFS on all the node without modifying the
old one (hybride: 1 DC EC2Snitch, 1 DC GPFS), again copying the EC2 DCs and
racks. Would this work ?

Also, Is this need of mirroring the exact DC and racks name mandatory in
those 2 procedures as I think it is ?

I think this worth a good testing anyway.

C*heers,

Alain

2015-08-03 22:04 GMT+02:00 John Wong <gokoproject@gmail.com>:

> Thanks Rob.
>
> But even if we do move out of EC2 in X years, unlike partitioner, I don't
> think there is much to do to switch snitch. Yes, I am evaluating if there
> is actually good benefit to use EC2.
>
> John
>
> On Mon, Aug 3, 2015 at 1:54 PM, Robert Coli <rcoli@eventbrite.com> wrote:
>
>> On Sun, Aug 2, 2015 at 6:22 PM, John Wong <gokoproject@gmail.com> wrote:
>>
>>> What other benefits can Ec2Snitch provide in a single-region, multi-az
>>> AWS deployment besides automatically setting dc and rack for you as the
>>> snitch reads from EC2 metadata.
>>> Obviously there is a concern with what if there is something wrong with
>>> the metadata server...
>>>
>>
>> GPFS is broadly similar to EC2Snitch/EC2MRSnitch, but is usable if you
>> for example have a DR site outside of EC2.
>>
>> In general I'd probably suggest using GPFS unless you're really sure
>> you'll always be in EC2, forever and always.
>>
>> =Rob
>>
>>
>
>

--001a1133a02e10c26a051dcf06f6
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi guys,<div><br></div><div>I am also wondering about this=
 kind of stuff, since we are going out of AWS and were using EC2Snitch. I f=
elt free to write it here, since it looks linked to previous comments.</div=
><div><br></div><div>@John, we have been using EC2Snitch for many years, an=
d it just works fine ! Plus, we don&#39;t have to keep the GPFS file up to =
date every time we add a node or remove one. EC2Snitch is quite strait-forw=
ard and easy to use. The first &quot;issue&quot; we have today, is indeed t=
o add a new DC from outside AWS. I am trying to find out how to do that rig=
ht now. If we figure out that this is easy, you can perfectly go with EC2Sn=
itch imho.</div><div><div><br></div><div>So what would be the procedure to =
go from EC2Snitch to GPFS ?</div><div><br></div><div>I see 2 possibilities =
and I am not sure of the result of each of these:</div><div><br></div><div>=
- Node by node, on a rolling restart way, and making sure that you use the =
former DC names and racks. Copying the EC2 network. Would this be doable or=
 would that need a full cluster restart at once ?</div><div>- Add the new D=
C using directly GPFS on all the node without modifying the old one (hybrid=
e: 1 DC EC2Snitch, 1 DC GPFS), again copying the EC2 DCs and racks. Would t=
his work ?</div><div><br></div><div>Also, Is this need of mirroring the exa=
ct DC and racks name mandatory in those 2 procedures as I think it is ?</di=
v></div><div><br></div><div>I think this worth a good testing anyway.</div>=
<div><br></div><div>C*heers,</div><div><br></div><div>Alain</div></div><div=
 class=3D"gmail_extra"><br><div class=3D"gmail_quote">2015-08-03 22:04 GMT+=
02:00 John Wong <span dir=3D"ltr">&lt;<a href=3D"mailto:gokoproject@gmail.c=
om" target=3D"_blank">gokoproject@gmail.com</a>&gt;</span>:<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">Thanks Rob.<div><br></div><div>But even=
 if we do move out of EC2 in X years, unlike partitioner, I don&#39;t think=
 there is much to do to switch snitch. Yes, I am evaluating if there is act=
ually good benefit to use EC2.</div><span class=3D"HOEnZb"><font color=3D"#=
888888"><div><br></div><div>John</div></font></span></div><div class=3D"HOE=
nZb"><div class=3D"h5"><div class=3D"gmail_extra"><br><div class=3D"gmail_q=
uote">On Mon, Aug 3, 2015 at 1:54 PM, Robert Coli <span dir=3D"ltr">&lt;<a =
href=3D"mailto:rcoli@eventbrite.com" target=3D"_blank">rcoli@eventbrite.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"><=
div class=3D"gmail_extra"><div class=3D"gmail_quote"><span>On Sun, Aug 2, 2=
015 at 6:22 PM, John Wong <span dir=3D"ltr">&lt;<a href=3D"mailto:gokoproje=
ct@gmail.com" target=3D"_blank">gokoproject@gmail.com</a>&gt;</span> wrote:=
<br><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-lef=
t:1px #ccc solid;padding-left:1ex"><div dir=3D"ltr"><div><div><div><div>Wha=
t other benefits can Ec2Snitch provide in a single-region, multi-az AWS dep=
loyment besides automatically setting dc and rack for you as the snitch rea=
ds from EC2 metadata.<br></div></div></div>Obviously there is a concern wit=
h what if there is something wrong with the metadata server...<br></div></d=
iv></blockquote><div><br></div></span><div>GPFS is broadly similar to EC2Sn=
itch/EC2MRSnitch, but is usable if you for example have a DR site outside o=
f EC2.</div><div><br></div><div>In general I&#39;d probably suggest using G=
PFS unless you&#39;re really sure you&#39;ll always be in EC2, forever and =
always.</div><div><br></div><div>=3DRob</div><div>=C2=A0</div></div></div><=
/div>
</blockquote></div><br></div>
</div></div></blockquote></div><br></div>

--001a1133a02e10c26a051dcf06f6--