T.R | Title | User | Personal Name | Date | Lines |
---|
1309.1 | Show directives (currently) do not take arguments | NANOVX::ROBERTS | Keith Roberts - DECmcc Toolkit Team | Mon Aug 05 1991 08:37 | 31 |
| >> What's missing? What's wrong? What have I missed?
>> How does the in_p look like if the default is passed along the MCC call i/f?
>> Will I find the symbol MCC_K_SEB_SYM_ARG_SHOW_ALL_ID ILV-coded or what?
(1) Show directives do not take arguments. So, you should not put any
in your MSL. Remove the 'REQUEST' stuff from the Show directive.
(2) For the Show directive you specified, you should expect:
Verb: SHOW
Entity: SEB <seb_entity> APPLICATION <appl_child_entity>
Partition: MCC_K_PRT_CHARACTERISTCS
In-P: null
For this directive you should return ** ALL ** your characteristic
attributes ... the FCL PM will pick out the attribute which was
requested on the command line and only display that one (not the
rest of the attributes you returned).
(3) You will not see the symbol MCC_K_SEB_SYM_ARG_SHOW_ALL_ID encoded
anywhere.
(4) Perhaps at a later date (DECmcc v2.0 ?) the Show directive may take
request arguments as you have done in your MSL, but not yet.
Does this help?
Keith Roberts
DECmcc Toolkit Team
|
1309.2 | what's next? | STKHLM::BERGGREN | Nils Berggren EIS/Project dpmt, Sweden DTN 876-8287 | Mon Aug 05 1991 10:08 | 40 |
| repl .1
>>> Does this help?
Both yes and no! I got the answer I was looking for, but I don't
like it. Why is this not documented (or if it is, so very hard
to find)? And furthermore, shouldn't the SAMPLE-AM provided be up to
date? I mean that in the MCC_SAMPLE_AM_SRVC_IF.MS there's a request
part for the show-directive, both for the global entity and for the
child entity (okey, in the show-directive routines the in_p is not
looked at, but anyway...).
If my entity whould have some 50-100 attributes (which will be the
case) I really don't think it's nice way of implementing the
show-directive to return all the 50-100 values and then let the PM
decide what to display. For my AM (and all other AMs as well) that
will result in a lot of unwanted net-trafic.
Shouldn't this have had higher priority so it would have been
implemented?
Is it documented somewhere what DECmcc v1.1 supports and more
important what is not supported but documented (and therefore thaught
as supported)?
/Nils
ps.
Sorry for being so negative, but my job is to implement an AM that
supports the GETEVENT and SHOW directive (later there will be more).
The first problem I ran into after a lot of trouble was the
event_report sent into the event_manager (MCC_EVENT_PUT) just going
down the drain. Reading the manuals about event-management, there's
nothing written about event-reports not being taken care of, just that
have to be provided!!!. At the moment this makes DECmcc not usable for
us. The second thing is this with the in_p and SHOW-directive and the
manual stating that the AM has to decode the in_p-argument....
What's next.....?
|
1309.3 | | TOOK::STRUTT | Management - the one word oxymoron | Mon Aug 05 1991 10:18 | 11 |
| Re: no argument to SHOW being passed down to the AM (yet).....
However, even though the AM does not get the argument passed down in
in_p (yet), do you agree that the correct thing happens as far as the
user is concerned, as the PM extracts the required attribute(s) from
the output arguments and only displays the requested attribute(s)?
I expect the release notes discuss which portions of the SRM are not
(yet) implemented completely, including the fact that the PMs do not
provide the "attributes wanted" argument.
Colin
|
1309.4 | On the Events problem | TOOK::STRUTT | Management - the one word oxymoron | Mon Aug 05 1991 10:37 | 23 |
| re: P.S. in .2 about events
Although this is slightly off the base topic, it seemed worthy of
reply. You make comment about mcc_event_put, and the events "going down
the drain". I don't follow *exactly* what you mean, but I guess that
you were unaware of the necessity of having a matching, previously set
up, mcc_event_get to ensure that the events you put get somewhere.
I checked the SRM, to be sure we had not missed out such an obvious (to
us) detail. On page 627, under the description for mcc_event_put, the
first line says:
This routine is used to send the event data to any threads
which have requested the matching event data.
On p369, in section 10.11, there is mention of a Getter thread and a
Putter thread, which is elaborated on in subsequent subsections of
10.11
If you feel there is vital information missing that causes
misunderstandings, please let me know (here, or via mail) and we will
address it in a future version of the SRM.
Colin
|
1309.5 | The PM does the right thing. | STKHLM::BERGGREN | Nils Berggren EIS/Project dpmt, Sweden DTN 876-8287 | Mon Aug 05 1991 11:10 | 11 |
| repl .3
>>> do you agree that the correct thing happens as far as the
>>> user is concerned, as the PM extracts the required attribute(s) from
>>> the output arguments and only displays the requested attribute(s)?
Yes, the PM should only show what I ask for. (but I still think its
the work of the AM to extract the info not the PM)
/Nils
|
1309.6 | event-stuff works, but... | STKHLM::BERGGREN | Nils Berggren EIS/Project dpmt, Sweden DTN 876-8287 | Mon Aug 05 1991 11:19 | 15 |
| repl .4
sorry about this, it's another story. I have the event-stuff working,
i.e. the icon goes red on the map and I can see that a particular rule
has fired. The only truble is that the info in the event-report sent
to the event-manager thru the MCC_EVENT_PUT and retreived in my
AM thru the MCC_EVENT_GET and from there sent to the FM that called my AM
via the out_p-argument can not be retreived via the PM nor in the
event-data file created by the alarms-fm, and I really need that info.
(what a sentence... hope you understand)
Thanks for your concern
/Nils
|
1309.7 | Show Directive Arguments - whats what? | NANOVX::ROBERTS | Keith Roberts - DECmcc Toolkit Team | Mon Aug 05 1991 11:36 | 47 |
| Nils,
There is some confusion with Show Directive Arguments.
o The SRM says in section 15.4.1.13 that Show takes no Arguments
o The SRM says in section 15.4.2 that Show takes optional arguments
>> This has been QAR'd today
- The Show directive CAN take an argument.
o The current FCL PM processes individual attribute requests by
filtering the entire partition of attributes returned by the MM,
and only displaying the requested attributes.
>> This was also QAR'd today
- The FCL should pass a list of Attributes for the Show directive
when specific attributes are requested in the command line. The MM
will process this attribute list, returning only the specified
attributes.
-----------
Q: How can you get your MM up and running quickly?
A: Today, even if you wrote code to process a list of Request Attributes,
the FCL will not encode the request into In-P. I suggest that you
postpone this functionality for when the FCL implements the Show
Request Argument (attributes_wanted).
I do not know when the FCL plans on implementing this functionality.
Q: The Sample AM has the Show Directive Request Arguments - how come?
A: Probably because the SRM templates have them. As you stated in .2
the actual source code does not attempt to read In-P to process the
request arguments. The YOURMM templates do not have the Request
Argument in the MSL - I will add this to the MSL for v1.2
I fully agree that returning many attribute values, when only one might have
been requested is inefficient. But as you can see, it is our intention to
correct this.
In a future release of the Design Framework, it is our hope to provide the
'guts' of the Set and Show directives; the protocol and overhead. You would
need only provide the code which interfaces to your entity; the conversion from
you entity's datatypes to MCC datatypes. The processing of the Show Arguments
would be done for you.
/keith
|
1309.8 | Does the GETEVENT directive from the FCL display your Evnet Report? | NANOVX::ROBERTS | Keith Roberts - DECmcc Toolkit Team | Mon Aug 05 1991 12:14 | 9 |
| Nils, if enter your GETEVENT directive from the FCL, do you see the
event report displayed correctly?
The Alarms software does not process event *arguments* .. only the Event
code (ie, the event name, like "counters zeroed"). So, you won't see
see any of the event report in anything the alarms code generates as
output.
/keith
|
1309.9 | works OK with FCL - not with IM-PM | STKHLM::BERGGREN | Nils Berggren EIS/Project dpmt, Sweden DTN 876-8287 | Tue Aug 06 1991 03:46 | 16 |
| repl .8
>>> Nils, if enter your GETEVENT directive from the FCL, do you see the
>>> event report displayed correctly?
Yes I do. Okey, I'm wrong again. The event report doesn't vanish in
this case, but (isn't there always a but...) the guys managing the
network will be using the iconic map PM, and there's no eventreport
displayed there when a rule has fired due to an event. Just that the
event has occured.
I know that this (notification FM not displaying event-reports) is not
yet supported, so I'll just have to wait...
thanks,
Nils
|