[Top] [All Lists]

Re: [nv-l] Disk and load monitoring on Linux and Windows Servers

To: nv-l@lists.us.ibm.com
Subject: Re: [nv-l] Disk and load monitoring on Linux and Windows Servers
From: Samuel Moñux <smonux@gmail.com>
Date: Thu, 26 May 2005 09:58:22 +0200
Delivery-date: Thu, 26 May 2005 08:58:59 +0100
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=k7j7/Zz1DV+Lmvzt2IiluiW9ThF5NeJr822jbUzwppYuFxRLm6rwb0a9iVjwjQqTofU64EMLEkHFL15vbQiDAdHUk+94/73kAsf/1qenDr0CcFcz9xB14OxDOeQvx50Iq18hdofAtPrL0MkX4rfKKZAxlCwi1etMwoY73ijbCnU=
Envelope-to: nv-l-archive@lists.skills-1st.co.uk
In-reply-to: <OF6670F588.A6570E9E-ON8525700C.007B49D7-8525700C.007DCD61@us.ibm.com>
References: <b052d0d0505250933423d17f7@mail.gmail.com> <OF6670F588.A6570E9E-ON8525700C.007B49D7-8525700C.007DCD61@us.ibm.com>
Reply-to: nv-l@lists.us.ibm.com
Sender: owner-nv-l@lists.us.ibm.com
Thanks for your advice. I understand now that the standard way is deal
with non-standardness.

The expression that I showed yesterday now works, and I've added
others which also work:
HDiskUtilUnix \
"dskPercent" \

ProcLivenessUnix \
"prErrorFlag" \

and more...

But now I need to get more info about what exactly crossed the
threshold. For example, I obtain messages like this:

1117091297 0  Thu May 26 09:08:17 2005 netview.carreras.sa       D
HDiskUtilUnix 4 threshold exceeded (>90.00): 100.00 HDiskUtilUnix 4

I don't know what exactly is that "partition 4" until I do a snmpwalk:

= STRING: /dev/pts

I would like to get a more detailed message on the event. How this can
be done? I need a way to attach another OID(usually a description
string) to the event.

Thanks, again.

2005/5/26, Leslie Clark <lclark@us.ibm.com>:
> You are off to a great start. Now you just need to find some mib variables
> that the servers you are monitoring can respond to. They will probably be
> proprietary. The MIB-II structure was designed with network equipment in
> mind, not servers. So there really isn't a standard way of providing that
> information. Well, there is a standard, and you tried that, but you know how
> server software vendors are. They have to be different, and try to become
> the defacto standard. I suspect that nagios was using a variety of different
> mib variables inside the various plugins.  No plugins here, but the people
> on this list can probably tell you which mibs they use for different kinds
> of servers. You should also search the archives of this list, because this
> subject comes up regularly. 
> http://lists.skills-1st.co.uk/mharc/html/nv-l/ 
> Some basic guidance: 
> The basic SNMP that comes with Microsoft operating systems won't know much
> besides its own name. When it is configured on the servers, the admin needs
> to 'check all the boxes' to get more of the MIB-II information (address
> tables, etc). Beyond that, most customers I've visited have something
> running on them, like Dell OpenManage, or IBM Director. These management
> agents have SNMP subagents. They know about RAID and all that. They even
> have some mibs in common. 
> In the Mib Browser, you can pull the MIB-II mgmt hosts branch and see if
> your intended target servers know anything about that. They might,or they
> might not. Then take a walk under the private branch. They may know a lot in
> that area if they have one of those agents on them. Next you will want to
> find the mib files that turn all of those dotted decimal identifiers to
> words, so you know what they are. The mib files are probably on the CDs that
> the management agents came on, and may even be loaded on the target servers
> somewhere. They are also generally available from mibdepot.com.  That's a
> really useful, if hard to navigate, website. So next you will be reading up
> on loading mibs. 
> You will find that different kinds of servers have different capabilities
> when it comes to reporting their status.  You might make good use of
> Smartsets to define groups of the same type. When you define data collection
> and thresholding, you can use one mib expression for one Smartset and
> another mib expression for another Smartset. 
>  Cordially,
>  Leslie A. Clark
>  IBM Global Services - Systems Mgmt & Networking
>  (248) 552-4968 Voicemail, Fax, Pager
>  Samuel Moñux <smonux@gmail.com> 
> Sent by: owner-nv-l@lists.us.ibm.com 
> 05/25/2005 12:33 PM 
> Please respond to
>  nv-l 
> To nv-l@lists.us.ibm.com 
> cc 
> Subject [nv-l] Disk and load monitoring on Linux and Windows Servers 
> Hi everyone,
>  I'm starting the deployment (planning the deployment to be more
>  accurate) of Tivoli Netview in an mixed environment of Linux, OpenBSD
>  and Windows servers, Cisco routers and switches, and RadioFrequency
>  terminals and antennas.
>  Due my lack of expertise with "true" NMS's, I'm trying to reach with
>  Netview the same level of functionality that Nagios has given us by
>  now, and in subsequent iterations I'll explore all the power of
>  Netview.
>  So, I have installed Tivoli Netview on an Suse Linux Enterprise Server
>  9, readed (most of) the Unix Administrator's guide and done some basic
>  setup.
>  Just now I'm trying to get notifications of high levels of load and
>  disk usage. Theorically, doesn't seem too difficult. I've added this
>  expression on mibExpr.conf
>  HDiskUtil \
>  "(hrStorageUsed / hrStorageSize) * 100"  \
>         . \
>         . / \
>         100 *
>  and using the Motif GUI, this to snmpCol.conf
>  MIB HDiskUtil HDiskUtil units EXPRESSION R
>  O Servidores_windows 60 90.000000 90.000000 A s 58720263 LIST 0
>  This doesn't work on Windows boxes and don't know why. But, it's
>  worse. Linux and OpenBSD boxes doesn't have hrStorageUsed at all on
>  their mib-2 tree. To get diskusage relevant OID's I have to look under
>  enterprises.ucdavis.
>  So, I'm a bit disoriented, Which is the "standard" way of deal with
>  net-snmp and Window implementations?. I'm specially interested on get
>  notifications of load, disk usage, and vital processes liveness.
>  Sorry if my question is too simplistic. I'll read any documentation
>  you point me at. I'm avid of detailed docs.
>  Thanks in advance,
>  Samuel

<Prev in Thread] Current Thread [Next in Thread>

Archive operated by Skills 1st Ltd

See also: The NetView Web