Zenoss Scalable setup

[jmp242]

jmp242 [http://community.zenoss.org/people/jmp242] created the discussion

"Re: Zenoss Scalable setup"

To view the discussion, visit: http://community.zenoss.org/message/68501#68501

--------------------------------------------------------------
GOM is a paid product from Zenoss and not available for Core that I know of. You could use distributed collectors, though I don't know if the zenpacks have been updated for v4 yet...

--
James Pulver
ZCA Member
LEPP Computer Group
Cornell University
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/68501#68501]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Fosiul Alam]

Fosiul Alam [http://community.zenoss.org/people/fosiul] created the discussion

"Re: Zenoss Scalable setup"

To view the discussion, visit: http://community.zenoss.org/message/68504#68504

--------------------------------------------------------------
Hi Thanks  for the reply.
So GOM is not an option

now with Distributed collection (which is a solution)

but the problem we are thinkign is :

Suppose we have 2 datacnetrer

DC1 and DC2
now we have cross connect betwen this 2 DC
and we have setup zenos in DC1 and using collector to collect event from DC2
Now if there is a problem in cross connect, we would not be able to get event from DC2

how will i overcome these problem ??

one solution is, setup 2 zenoss in both DC
have their own instances their, but then we need to look at 10 DC individually !!

but by using distributed collector :
can the event be collect from Dc1 and DC2 in both server ?? (Can the event be stored in both DC1 and DC2??)

So all the event will come to DC1 also event will be in dc2, so in event of cross conect problem we dont have any issue ??

can it be done ??
Does it make sense ?? if not please tel me, i will try to explain.
Thanks
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/68504#68504]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[jmp242]

jmp242 [http://community.zenoss.org/people/jmp242] created the discussion

"Re: Zenoss Scalable setup"

To view the discussion, visit: http://community.zenoss.org/message/68505#68505

--------------------------------------------------------------
Well, in the sense that Core is open source software, you probably could code something to do what you want, but it would likely be a large project. The ZCA and the community in general have some skilled Zenoss consultants / coders, but it would probably cost a bit to get something like this written as a zenpack...

--
James Pulver
ZCA Member
LEPP Computer Group
Cornell University
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/68505#68505]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Fosiul Alam]

Fosiul Alam [http://community.zenoss.org/people/fosiul] created the discussion

"Re: Zenoss Scalable setup"

To view the discussion, visit: http://community.zenoss.org/message/68507#68507

--------------------------------------------------------------
Hi
Thanks for quick reply.

just few confusion..

with my previous qustion ..

In event of losing connectivity between data centers and assuming we have core in dc1  and collector in dc2. Would collector in dc2 continue monitoring devices, send notifictions and run zenactions ?

Please let me know
Thanks
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/68507#68507]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[jmp242]

jmp242 [http://community.zenoss.org/people/jmp242] created the discussion

"Re: Zenoss Scalable setup"

To view the discussion, visit: http://community.zenoss.org/message/68509#68509

--------------------------------------------------------------
I can't say that I know, I don't run distributed collectors. I expect notifications etc are processed on the master, not on collectors.

--
James Pulver
ZCA Member
LEPP Computer Group
Cornell University
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/68509#68509]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[tcaiazza]

tcaiazza [http://community.zenoss.org/people/tcaiazza] created the discussion

"Re: Zenoss Scalable setup"

To view the discussion, visit: http://community.zenoss.org/message/68512#68512

--------------------------------------------------------------
Fosiul,

Here is my understanding of how things work, someone please correct me if I'm wrong.  If you lose connectivity between the zenhub(the server that centrally manages everything) and a collector, the collector continues to get alerts but will not send them back to the zenhub.  From the zenhub alerts are sent out with zenactions, so you will not see any alerts that the collector receives.  But you will see alerts for issues connecting to the collector, which should tip you off to any issues.  Once a connection has been re established with zenhub and the collector, the collector will update zenhub with all the alerts it has received.  At this point then you should get your alerts.

I have one zenhub server, with 3 remote collectors monitoring 10+ datacenters.  In the case when one collector has been cut of from the zenhub, that remote collector continues to receive snmptraps and syslog messages.  But again we don't see any of these alerts, until the connection between the zenhub and the collector has been re-established.

I've found this to be fine because our datacenters being cut off from each other can only happen in the case of multiple failures.
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/68512#68512]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[David Carmean]

David Carmean [http://community.zenoss.org/people/dcarmean] created the discussion

"Re: Zenoss Scalable setup"

To view the discussion, visit: http://community.zenoss.org/message/68520#68520

--------------------------------------------------------------
Let me clarify some terminology.

* Event1. An object created inside the remote collector daemon (which in the case of zensyslog is a translation of an external message) that is sent across the network to the zenhub process.
2. A representation of this object that is stored in the MySQL tables and visible as a row in the Event Console.

* Alert:* 1: Now called Notifications in Zenoss 4.x
* The external result of zenactiond doing something based on an event filter (now called Event triggers), like sending an email or an sms


What you are discussing is the queueing of generated Events on the remote collector while connectivity to the zenhub daemon is interrupted.  Yes, each individual remote daemon (zenperfsnmp, zenping, etc) will store up to 5000 (by default, adjustable) events in memory if the connection to the upstream zenhub is lost.  Once that 5000 is reached, the oldest events are dropped.
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/68520#68520]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]