Which functions are lost when using NVSP without NV/6000?
ANSWER
The way to view the relationship is that NV Service Point provides
some services that NV/6000 uses. NV Service Point provides a set
of APIs, so that an application may:
1. Send alerts to host NetView
2. Receive RUNCMDs from host NetView, and return responses
3. Send messages to operators on host NetView
4. Control devices attached to async interfaces on the RISC
System/6000
5. Send and receive files to/from CICS DDM
NetView/6000 provides applications to exploit the first two of these
APIs: tralertd, which converts NV6000 events into alerts, and spappld
which waits for and executes commands sent as RUNCMDs.
Without NV/6000 you need to have other applications to use the APIs,
which may be provided by a vendor. These may be hardware-specific
applications or "generic" applications such as Diederich's NetScript,
or you may use the sample code that comes with NV Service Point, or
the code published in ITSO redbooks (Experiences Using AIX NetView
Service Point, GG24-3701) or via the examples in the recent redbook:
"Examples of Using NetView for AIX", GG24-4327.
QUESTION
Let me make the background of the question clearer:
Back when NM/6000 still existed NM/6000 was the only application needed
to convert alertable errlog entries to NMVTs and send them to the host.
In my opinion NV/SP should have the same function.
From your answer I conclude that I need program besides NV/SP to
have the same function as I had when using NM/6000.
Is this correct ?
ANSWER
Yes you are correct. The process for handling AIX 'alertable' errlog
reports that NM/6000 provided has become part of the NV/6000 SNMP
subagent - trapgend. Where NM/6000 converted errlog entries into
alerts and forwarded them to the host, trapgend converts them into
SNMP traps. The tralertd process I referred to above can then convert
the trap into an alert, and use NetView Service Point to forward it.
Clearly, if you do not have much use of TCP/IP, or do not want an
SNMP manager in the network, this is not ideal. You could write a
NV Service Point application to register for alertable errlog reports
and create alerts from them (ie replicating the function in NM/6000).
Alternatively there will be developments in SNA support on the RISC
System/6000 in the near future which will provide another alternative.
QUESTION
This is exactly the position I (and a customer I know of) are in:
no noticable TCP/IP network, but still wanting 'alertable' errlog
entries
to be sent to the host.
To prevent me reinventing the wheel:
Is there information (or an example) to start writing the routine you
described ??
ANSWER
The redbooks referred to above have examples for the 'second' part of
the process, ie creating and sending an alert. Triggering this
automatically requires you to add an object to the Error Notification
object class. This can be a program or shell script, and it can
receive error id information so that it may extract all the error
detail.
I do not know of any sample for this, but I have just tried a playing
with it and it looks straight-forward:
1. Export the existing profiles, command: odmget errnotify > yourfile
2. Add an entry for your process. It will look something like:
errnotify:
en_pid = 0
en_name = "trap-notify"
en_persistenceflg = 1
en_label = ""
en_crcid = 0
en_class = ""
en_type = ""
en_alertflg = "TRUE"
en_resource = ""
en_rtype = ""
en_rclass = ""
en_method = "/your/shell/or/program/file"
3. Remove all the error notification entries: odmdelete -o errnotify
4. Re-add them all, plus your new one: odmadd < yourfile
Look in Infoexplorer for more detail about these commands.
S e a r c h - k e y w o r d s:
NETVIEW SERVICE POINT AIX NETVIEW/6000 ERRLOG ALERT
WWQA: ITEM: RTA000064485 ITEM: RTA000064485
Dated: 10/1995 Category: ITSCSAIXNV6
This HTML file was generated 99/06/24~12:43:23
Comments or suggestions?
Contact us