|
More questions about REPORTER
Well, I was able to fix the above problem. For those of you that are
not DATATRIEVE programmers, you need to put an ';' at the end of the
line, just like in the documentation... 8^).
I do have a more general REPORTER question. I'm afraid I'm going to
have to ramble, because it's more theory, than just a black and white
answer...
Here's the scenerio:
I have 3 domains that I want to use REPORTER for. The types of entities
are mostly NODE4s; I've got DECrouters, and VAX workstations. I also
have 2-3 LANbridge 200s that I want to report on.
I have setup the EXPORTER, to export the COUNTERS partition for all of
the above entities. When I tried to setup my CDD database files, I
followed the documentation but was puzzled by DTR/CDD. I don't code in
DTR, so I'm quite naive...smile.
The first time I setup my CDD$DEFAULT logical, I put, CDD$TOP.REPORTS.
I pointed this to the Rdb database which was one of my EXPORT Targets.
I thought I had setup my MCC_EXPORT logical properly, and then I ran
the 3 Phase IV procedures, found in MCC_REPORTS_FILES. It looked like
this was fine...
After I ran the 3 Phase IV procedures, found in MCC_REPORTS_FILES to
update my dictionary. This is where I get confused...
I assumed that I would need to repeat the above step for "each domain"
that I wanted to report on? I also *assumed* that I needed to create a
new CDD$DEFAULT, since the documentation says that if you want to point
to different databases you need to change this default... So the 2
subsequent steps I took, I defined CDD$DEFAULT to CDD$TOP.REPORTS2, and
then CDD$TOP.REPORTS3... I got several informational messages from
DTR, but it looks like the dictionary was updated with the
procedures...
Well, when I tried to run REPORTS, I got the following error in my
logfile:
� MCC_RPTS_NODE4_LINE_TRAFFIC_DTR.LOG ,
Element "MCC_EXPORT" not found in dicitionary
"CURRENT" is not a readied source, collection, or list
"CURRENT" is not a readied source, collection, or list
"CURRENT" is not a readied source, collection, or list
and the job bombs... HELP! Please send mail to JUNG::GALVIN or answer
here. Thanks.
/Mic
|
|
Mic,
I'm not a Report person but it seems to me that if you want
"node4 line traffic" report you have to export not only
node4 entity but also node4 line entity (or only node4 line ?).
Please, check Report documentation.
Sam
BTW,
> I have setup the EXPORTER, to export the COUNTERS partition for all of
> the above entities.
This is impossible. Exporter exports ALL PARTITIONS for a target entity.
Regards,
|
| Sam,
I quite confused by your last advise... I was under the impression
that it goes like this...
1.) Setup HISTORIAN background process and then use RECORD for an
entity. I used COUNTERS, on NODE4 'foo'.
2.) Setup EXPORTER background process and then setup EXPORTing of
an entity. In otherwords, now I'm EXPORTING the COUNTERS data from
node 'foo', right?
3.) I setup the REPORTER, by setting up the MCC_EXPORT logical in
DTR and then running the menu option for the target entity, 'foo',
right?
I assumed that if you RECORD COUNTERS and EXPORT *everything*, the
report should include all of the child entities of the global entity,
right? In otherwords, I would hope that if I wanted to get COUNTERS
information on a NODE4, I should get any of the CHILD ENTITY
information... Please set me straight.
BTW, I'm really puzzled by the REPORTER setup and need some help.
Please send mail to JUNG::GALVIN, or call (201) 368-5501. Thanks.
/Mic
|
|
> 1.) Setup HISTORIAN background process and then use RECORD for an
> entity. I used COUNTERS, on NODE4 'foo'.
>
> 2.) Setup EXPORTER background process and then setup EXPORTing of
> an entity. In otherwords, now I'm EXPORTING the COUNTERS data from
> node 'foo', right?
This is NOT right.
Firstly, you have to understand that Historian and Exporter are
INDEPENDENT FMs. When you set up recording and current time
exporting they do not know about each other and do not communicate
in any case.
Secondly, Exporter works on per entity bases. It means that ALL
attribute defined in the Data Dictionary for a particular entity
class are written into RDB (except attribute whose data types are
supported by the Exporter).
Thirdly, when you set up PAST TIME exporting the data flow is as
follow ( it's described in SRM):
- Exporter makes a call to get past time data;
- Information manager tries to satisfy this request using
historical data priviously recorded by the Historian;
- If information manager can not find historical data for
a specified period of time it make a request to the
appropriate AM. The current implementation of AM's does
not have any knowledge about past time data.
So in the situation when you record, for example, only counters
and later use Past Time Export in the Rdb file you have ONLY
counters and nothing more which is not enough for reports.
> I assumed that if you RECORD COUNTERS and EXPORT *everything*, the
> report should include all of the child entities of the global entity,
> right? In otherwords, I would hope that if I wanted to get COUNTERS
> information on a NODE4, I should get any of the CHILD ENTITY
> information... Please set me straight.
This is wrong.
I already explained that current Export and Record are not relaited to
other. Also now it's clear that if record only counters and use Past
Time Export in the RDB you have ONLY counters.
Reports uses RDB as the only source to build a particular report.
If RDB contains information only about NODE4 how and where Report can
get information about child entities.
In the current version of MCC it's the user's responsibility to export
all information required for producing reports. I believe that in the
report manual there is a list of entity classes/subclasses required
for a particular report.
SAm
|