Send Netdot-users mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
https://osl.uoregon.edu/mailman/listinfo/netdot-users
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Netdot-users digest..."
Today's Topics:
1. server sizing (Newton, Terry M)
2. Re: server sizing (Carlos Vicente)
3. Re: server sizing (Newton, Terry M)
4. Re: server sizing (Carlos Vicente)
5. Re: server sizing (Phil Regnauld)
6. Re: server sizing (William Bulley)
7. Re: server sizing (Phil Regnauld)
----------------------------------------------------------------------
Message: 1
Date: Fri, 14 Jun 2013 13:33:01 +0000
From: "Newton, Terry M" <[email protected]>
Subject: [Netdot-users] server sizing
To: "'[email protected]'" <[email protected]>
Message-ID:
<[email protected]>
Content-Type: text/plain; charset="us-ascii"
Hello, new user here and I have some questions about sizing out a server or
servers for monitoring a large enterprise network. I am just beginning my
deployment of netdot and am using a virtual machine. I will eventually be
monitoring 60 routers and 600 forty-eight port switches. I am probably 15%
into the project and am beginning to see a large 10 to 15 minute spike in
server resources at the top of the hour whenever netdot does it's arp and
forwarding table poll.
For a network of this size, can somebody please point me at some specifications
for sizing the server? Will a vm suffice for this? I am thinking not. Will
one server be enough? Any advice is greatly appreciated!!!! Thanks!
Confidentiality Notice: This e-mail message, including any attachments,
is for the sole use of the intended recipient(s) and may contain
confidential and privileged information. Any unauthorized review,
use, disclosure or distribution is prohibited. If you are not the
intended recipient, please contact the sender by reply
e-mail and destroy all copies of the original message.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://osl.uoregon.edu/pipermail/netdot-users/attachments/20130614/71141d1c/attachment-0001.html
------------------------------
Message: 2
Date: Fri, 14 Jun 2013 10:08:25 -0400
From: Carlos Vicente <[email protected]>
Subject: Re: [Netdot-users] server sizing
To: "Newton, Terry M" <[email protected]>
Cc: "'[email protected]'" <[email protected]>
Message-ID: <[email protected]>
Content-Type: text/plain; charset=windows-1252
Hello Terry,
I won't comment on specific hardware specs for fear of being off. But I
can mention a couple of things that will help:
* Don't do full device updates (-I parameter) every hour unless really
necessary. You could, for example, collect ARP (-A) and forwarding
tables (-F) every hour, and do full updates (-I) once or twice a day.
This is easy to do by playing a bit with cron jobs.
* Separate the application into two separate physical machines. One that
runs the database and the web server, and another one that does the
polling. Whether you should run these as VMs or not is irrelevant. What
matters is what resources are allocated to each VM. The downside of this
approach is that you have to maintain the code on both servers.
I hope this helps.
cv
On 6/14/13 9:33 AM, Newton, Terry M wrote:
>
> Hello, new user here and I have some questions about sizing out a
> server or servers for monitoring a large enterprise network. I am
> just beginning my deployment of netdot and am using a virtual
> machine. I will eventually be monitoring 60 routers and 600
> forty-eight port switches. I am probably 15% into the project and am
> beginning to see a large 10 to 15 minute spike in server resources at
> the top of the hour whenever netdot does it?s arp and forwarding
> table poll.
>
>
>
> For a network of this size, can somebody please point me at some
> specifications for sizing the server? Will a vm suffice for this?
> I am thinking not. Will one server be enough? Any advice is greatly
> appreciated!!!! Thanks!
>
> Confidentiality Notice: This e-mail message, including any
> attachments, is for the sole use of the intended recipient(s) and may
> contain confidential and privileged information. Any unauthorized
> review, use, disclosure or distribution is prohibited. If you are
> not the intended recipient, please contact the sender by reply e-mail
> and destroy all copies of the original message.
>
>
>
> _______________________________________________ Netdot-users mailing
> list [email protected]
> https://osl.uoregon.edu/mailman/listinfo/netdot-users
------------------------------
Message: 3
Date: Fri, 14 Jun 2013 14:14:10 +0000
From: "Newton, Terry M" <[email protected]>
Subject: Re: [Netdot-users] server sizing
To: "'Carlos Vicente'" <[email protected]>
Cc: "'[email protected]'" <[email protected]>
Message-ID:
<[email protected]>
Content-Type: text/plain; charset="us-ascii"
How about setting up polling per subnet at different intervals? Is there a way
to do that?
-----Original Message-----
From: Carlos Vicente [mailto:[email protected]]
Sent: Friday, June 14, 2013 9:08 AM
To: Newton, Terry M
Cc: '[email protected]'
Subject: Re: [Netdot-users] server sizing
Hello Terry,
I won't comment on specific hardware specs for fear of being off. But I can
mention a couple of things that will help:
* Don't do full device updates (-I parameter) every hour unless really
necessary. You could, for example, collect ARP (-A) and forwarding tables (-F)
every hour, and do full updates (-I) once or twice a day.
This is easy to do by playing a bit with cron jobs.
* Separate the application into two separate physical machines. One that runs
the database and the web server, and another one that does the polling. Whether
you should run these as VMs or not is irrelevant. What matters is what
resources are allocated to each VM. The downside of this approach is that you
have to maintain the code on both servers.
I hope this helps.
cv
On 6/14/13 9:33 AM, Newton, Terry M wrote:
>
> Hello, new user here and I have some questions about sizing out a
> server or servers for monitoring a large enterprise network. I am just
> beginning my deployment of netdot and am using a virtual machine. I
> will eventually be monitoring 60 routers and 600 forty-eight port
> switches. I am probably 15% into the project and am beginning to see a
> large 10 to 15 minute spike in server resources at the top of the hour
> whenever netdot does it's arp and forwarding table poll.
>
>
>
> For a network of this size, can somebody please point me at some
> specifications for sizing the server? Will a vm suffice for this?
> I am thinking not. Will one server be enough? Any advice is greatly
> appreciated!!!! Thanks!
>
> Confidentiality Notice: This e-mail message, including any
> attachments, is for the sole use of the intended recipient(s) and may
> contain confidential and privileged information. Any unauthorized
> review, use, disclosure or distribution is prohibited. If you are not
> the intended recipient, please contact the sender by reply e-mail and
> destroy all copies of the original message.
>
>
>
> _______________________________________________ Netdot-users mailing
> list [email protected]
> https://osl.uoregon.edu/mailman/listinfo/netdot-users
Confidentiality Notice: This e-mail message, including any attachments,
is for the sole use of the intended recipient(s) and may contain
confidential and privileged information. Any unauthorized review,
use, disclosure or distribution is prohibited. If you are not the
intended recipient, please contact the sender by reply
e-mail and destroy all copies of the original message.
------------------------------
Message: 4
Date: Fri, 14 Jun 2013 10:17:27 -0400
From: Carlos Vicente <[email protected]>
Subject: Re: [Netdot-users] server sizing
To: "Newton, Terry M" <[email protected]>
Cc: "'[email protected]'" <[email protected]>
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1
Not per subnet.
But you can use the following parameter to select groups of devices per
name:
-m, --matching <regex> Update only devices whose names
match pattern (with -B, -D, -E)
cv
On 6/14/13 10:14 AM, Newton, Terry M wrote:
> How about setting up polling per subnet at different intervals? Is there a
> way to do that?
>
> -----Original Message-----
> From: Carlos Vicente [mailto:[email protected]]
> Sent: Friday, June 14, 2013 9:08 AM
> To: Newton, Terry M
> Cc: '[email protected]'
> Subject: Re: [Netdot-users] server sizing
>
> Hello Terry,
>
> I won't comment on specific hardware specs for fear of being off. But I can
> mention a couple of things that will help:
>
> * Don't do full device updates (-I parameter) every hour unless really
> necessary. You could, for example, collect ARP (-A) and forwarding tables
> (-F) every hour, and do full updates (-I) once or twice a day.
> This is easy to do by playing a bit with cron jobs.
> * Separate the application into two separate physical machines. One that runs
> the database and the web server, and another one that does the polling.
> Whether you should run these as VMs or not is irrelevant. What matters is
> what resources are allocated to each VM. The downside of this approach is
> that you have to maintain the code on both servers.
>
> I hope this helps.
>
> cv
>
> On 6/14/13 9:33 AM, Newton, Terry M wrote:
>> Hello, new user here and I have some questions about sizing out a
>> server or servers for monitoring a large enterprise network. I am just
>> beginning my deployment of netdot and am using a virtual machine. I
>> will eventually be monitoring 60 routers and 600 forty-eight port
>> switches. I am probably 15% into the project and am beginning to see a
>> large 10 to 15 minute spike in server resources at the top of the hour
>> whenever netdot does it's arp and forwarding table poll.
>>
>>
>>
>> For a network of this size, can somebody please point me at some
>> specifications for sizing the server? Will a vm suffice for this?
>> I am thinking not. Will one server be enough? Any advice is greatly
>> appreciated!!!! Thanks!
>>
>> Confidentiality Notice: This e-mail message, including any
>> attachments, is for the sole use of the intended recipient(s) and may
>> contain confidential and privileged information. Any unauthorized
>> review, use, disclosure or distribution is prohibited. If you are not
>> the intended recipient, please contact the sender by reply e-mail and
>> destroy all copies of the original message.
>>
>>
>>
>> _______________________________________________ Netdot-users mailing
>> list [email protected]
>> https://osl.uoregon.edu/mailman/listinfo/netdot-users
>
> Confidentiality Notice: This e-mail message, including any attachments,
> is for the sole use of the intended recipient(s) and may contain
> confidential and privileged information. Any unauthorized review,
> use, disclosure or distribution is prohibited. If you are not the
> intended recipient, please contact the sender by reply
> e-mail and destroy all copies of the original message.
>
------------------------------
Message: 5
Date: Fri, 14 Jun 2013 16:28:53 +0200
From: Phil Regnauld <[email protected]>
Subject: Re: [Netdot-users] server sizing
To: Carlos Vicente <[email protected]>
Cc: "Newton, Terry M" <[email protected]>,
"'[email protected]'" <[email protected]>
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii
Carlos Vicente (cvicente.uo) writes:
>
> -m, --matching <regex> Update only devices whose names
> match pattern (with -B, -D, -E)
This is extremely useful when monitoring large networks with
standardized
naming schemes. For examples, I've worked on sites where the naming
conventions are:
[r|s|srv][country][city][sequence]
For example:
rdkcph001
With such a scheme it's easy to only run updates for switches in a
particular POP or region, or even matching odd numbered devices one
hour and even numbered devices the other.
Cheers,
PHil
------------------------------
Message: 6
Date: Fri, 14 Jun 2013 10:38:11 -0400
From: William Bulley <[email protected]>
Subject: Re: [Netdot-users] server sizing
To: Carlos Vicente <[email protected]>
Cc: "Newton, Terry M" <[email protected]>,
"'[email protected]'" <[email protected]>
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii
According to Carlos Vicente <[email protected]> on Fri, 06/14/13 at 10:08:
>
> * Separate the application into two separate physical machines. One that
> runs the database and the web server, and another one that does the
> polling.
I would like to ask you to clarify your above advice just a tad.
I understand the concept of running two instances of the application,
but does the second server (the one "that does the polling") run a
second instance of "the database"? If not, do you imply some version
of data transport of database writes back to the first server.
On the other hand, if yes, won't that imply there is a need for some
sort of database merging, or syncing, of the polling results with the
ARP and other data stored on the first server's database? Thanks!
Regards,
web...
--
William Bulley Email: [email protected]
72 characters width template ----------------------------------------->|
------------------------------
Message: 7
Date: Fri, 14 Jun 2013 16:56:16 +0200
From: Phil Regnauld <[email protected]>
Subject: Re: [Netdot-users] server sizing
To: William Bulley <[email protected]>
Cc: Carlos Vicente <[email protected]>,
"'[email protected]'" <[email protected]>,
"Newton, Terry M" <[email protected]>
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii
William Bulley (web) writes:
>
> I understand the concept of running two instances of the application,
> but does the second server (the one "that does the polling") run a
> second instance of "the database"? If not, do you imply some version
> of data transport of database writes back to the first server.
The netdot tools on server "poller" are simply configured to talk to
the DB
on the DB server "db".
> On the other hand, if yes, won't that imply there is a need for some
> sort of database merging, or syncing, of the polling results with the
> ARP and other data stored on the first server's database? Thanks!
No need.
------------------------------
_______________________________________________
Netdot-users mailing list
[email protected]
https://osl.uoregon.edu/mailman/listinfo/netdot-users
End of Netdot-users Digest, Vol 55, Issue 4
*******************************************
