Expanding on Jeffrey's comment on 'nvdbformat" and its uses,
here is an example that uses nvdbformat to create an html-page
for all routers in your ovw database.
If the file is called "routers.html.conf", then the command to create
the output would be "nvdbformat -f routers.html.conf".
Its actually quite cool, in that it makes creating different slices of
the database quite easy.
Regards,
Gary Boyles, Intel
One Note: I don't know if this will back-rev to a NetView NT V5.0a
system, so if it doesn't work, you'll have to upgrade.
Here is the file
-------------------- Cut Here --------------------
SELECTFIELD:1:Selection Name
SELECTFIELD:2:IP Status
SELECTFIELD:3:SNMP sysObjectID
SELECTFIELD:4:SNMP sysContact
SELECTFIELD:5:SNMP sysLocation
SELECTRULE:isNode=TRUE
#SELECTRULE:isIPRouter=TRUE
HEADER:<HTML>
HEADER:<HEAD>
HEADER:<BOLD>
HEADER:<TITLE>All_IP_Routers Table</TITLE>
HEADER:</BOLD>
HEADER:</HEAD>
HEADER:<TABLE border=3 align=center>
HEADER:<CAPTION align=top>
HEADER:<H3>List of Nodes Monitored as Routers<H3>
HEADER:</CAPTION>
HEADER:<FONT SIZE=2>
HEADER:<TR>
HEADER:<H3>
HEADER:<TH>Node Name</TH>
HEADER:<TH>IP Status</TH>
HEADER:<TH>SNMP sysObjectID</TH>
HEADER:<TH>SNMP sysContact</TH>
HEADER:<TH>SNMP sysLocation</TH>
HEADER:</H3>
HEADER:</TR>
HEADER:</FONT>
OUTPUT:<TR>
OUTPUT:<TH>${1}</TH>
OUTPUT:<TH>${2}</TH>
OUTPUT:<TH>${3}</TH>
OUTPUT:<TH>${4}</TH>
OUTPUT:<TH>${5}</TH>
OUTPUT:</TR>
FOOTER:</TABLE>
FOOTER:<BR>
FOOTER:<BR>
FOOTER:</HTML>
-------------------- Cut Here --------------------
> -----Original Message-----
> From: Jeffrey_Snover AT TIVOLI DOT COM [SMTP:Jeffrey_Snover AT TIVOLI DOT COM]
> Sent: Friday, August 21, 1998 7:26 AM
> To: NV-L AT UCSBVM.UCSB DOT EDU
> Subject: Re: NetView/NT Database useage
>
> > I'm using MSSQL database and connectivity to the database is working
> > the events come from the database and snmpcollect also write info
> into
> > the database, then why do I still need ovwdb?
>
>
> >>Using a relational database does not replace the one maintained by
> ovwdb.
> >>There are three databases (object, topology, and map) and unless they
> >>have really changed the NT product in ways I don't understand (which is
> >>possible, I'll admit, since I don't work on it), only topology would get
> >>loaded into the relational db. This is how it works on UNIX.
>
> The NetView/NT product uses a relational DB in the following areas:
> Native Store for Events
> Native Store for SNMP Historical data
> Native Store for Availability history (e.g. that node was down
> 7 times yesterday)
> Exported Topology Data
> Exported OVWDB Data
> (Via the NVDBEXPORT command which allows you to specify
> which objects to export and which fields should go into
> which tables. This is pretty outrageously wonderful, so
> give it a whirl and hook up a report writer like Crystal
> Reports or Access to the exported tables and you'll be
> rocking.
>
> If you are new to NVDBEXPORT that probably means you
> haven't checked out NVDBIMPORT or NVDBFORMAT. NVDBIMPORT
> allows you to do a bulk load into the database along the
> lines of a mail merge paradigm:
> hostname,field-to-load1,field-to-load2
> foo.dev.com,TRUE,\\foo
> bar.dev.com,TRUE,\\bar
>
> NVDBFORMAT is sortof a OVWDB-oriented printf() that
> overdosed on steriods. [It is wonderful - really!] You
> tell it which objects you want to operate on, what fields
> you want and then a series of printf-like statements you
> want generated (you can also specify headers and footers
> and in V5.1 you get sorting, grouping, group-headers,
> and group-footers). Try this out - you'll love it.
>
> UNIX users will get NVDBIMPORT and NVDBFORMAT in the
> 5.1 release.
> )
>
> Without going into tons of detail, the reason why you still have
> a non-relational topology and object database is just pure
> pragmatics. The data usage model is such that we get GREAT
> performance on Events, Collected data, and Availability but not
> on the others.
>
> Jeffrey Snover
> Tivoli
|
