Mailing-List: contact dev-help@stratos.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@stratos.apache.org
Received-SPF: pass (athena.apache.org: domain of reka@wso2.com designates
 209.85.216.48 as permitted sender)
MIME-Version: 1.0
From: Reka Thirunavukkarasu <reka@wso2.com>
Date: Sun, 14 Dec 2014 10:07:47 +0530
Message-ID: 
 <CAFhGutL-cEx8UZSBKBi-NAD5ym=G_LOycfZypcyAnGqfDiHT0A@mail.gmail.com>
Subject: Executing cluster monitor periodically as well as on demand
To: dev <dev@stratos.apache.org>
Cc: Lakmal Warusawithana <lakmal@wso2.com>, Nirmal Fernando <nirmal@wso2.com>,
	Lahiru Sandaruwan <lahirus@wso2.com>, Imesh Gunaratne <imesh@wso2.com>
Content-Type: multipart/alternative; boundary=001a113a74ea471f9f050a25b18b

--001a113a74ea471f9f050a25b18b
Content-Type: text/plain; charset=UTF-8

Hi All,

As we now support adding cluster instances dynamically, we will have to
wait for next one minute to get the members started in that particular
cluster instance which will delay other dependencies startup when it comes
to scaling or terminate and start again. Also, after adding the members in
obsolete member list when terminating a cluster, we will again have to wait
for one minute to get the members terminated. This will delay the action
based on the termination specially when it comes to terminate-dependent or
terminate-all scenarios.

Can we do $subject by introducing locking support for the cluster monitor?
So that whenever needed, we can directly execute the cluster monitor by
taking lock to take decision only for min check and obsolete check. Let
Cluster monitor periodically execute to take decision like min check,
obsolete check and scaling. Because, AFAIK for scaling we will have to wait
minimum one minute to get all the latest stats.

Thanks,
Reka

-- 
Reka Thirunavukkarasu
Senior Software Engineer,
WSO2, Inc.:http://wso2.com,
Mobile: +94776442007

--001a113a74ea471f9f050a25b18b
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi All,<div><br></div><div>As we now support adding cluste=
r instances dynamically, we will have to wait for next one minute to get th=
e members started in that particular cluster instance which will delay othe=
r dependencies startup when it comes to scaling or terminate and start agai=
n. Also, after adding the members in obsolete member list when terminating =
a cluster, we will again have to wait for one minute to get the members ter=
minated. This will delay the action based on the termination specially when=
 it comes to terminate-dependent or terminate-all scenarios.</div><div><br>=
</div><div>Can we do $subject by introducing locking support for the cluste=
r monitor? So that whenever needed, we can directly execute the cluster mon=
itor by taking lock to take decision only for min check and obsolete check.=
 Let Cluster monitor periodically execute to take decision like min check, =
obsolete check and scaling. Because, AFAIK for scaling we will have to wait=
 minimum one minute to get all the latest stats.</div><div><br></div><div>T=
hanks,</div><div>Reka<br clear=3D"all"><div><br></div>-- <br><div class=3D"=
gmail_signature"><div dir=3D"ltr"><div>Reka Thirunavukkarasu<br>Senior Soft=
ware Engineer,<br>WSO2, Inc.:<a href=3D"http://wso2.com" target=3D"_blank">=
http://wso2.com</a>,<br></div>Mobile: +94776442007<br><div><br><span><font =
color=3D"#888888"><span style=3D"font-size:13px"><span style=3D"border-coll=
apse:collapse;color:rgb(80,0,80)"><font face=3D"arial, helvetica, sans-seri=
f"><span style=3D"font-family:arial,sans-serif;font-size:13px;color:rgb(0,0=
,0)"></span></font></span></span></font></span><span><font color=3D"#888888=
"><span style=3D"font-size:x-small"><br></span></font></span></div></div></=
div>
</div></div>

--001a113a74ea471f9f050a25b18b--