T.R | Title | User | Personal Name | Date | Lines |
---|
2306.1 | | YAHEY::BOSE | | Mon Feb 10 1992 09:57 | 15 |
|
>>As you all know, is the SNMP AM only sends hexcode for the moste statusfields.
The SNMP AM can display both ASCII and hex strings. (ifDescr is an
example of the former, while ifPhyAddress is an example of the
latter). Which status attribute are you talking about ?
In a private mib, if an object is of type OCTET STRING, then it is
displayed by the SNMP AM as a hex string. If its type is DisplayString,
then it is displayed as an ASCII string (human-readable).
I guess I don't understand what exact problem you are facing. Could
you please elaborate.
rahul
|
2306.2 | Just a Translate | ZUR01::SCHNEIDERR | | Tue Feb 11 1992 03:39 | 17 |
| Rahul,
Lets explain, what I'm looking for:
If in a privat MIB an object is of type OCTET STRING, than we have to look into
the MIB file for the explaination of this octet, to kinow what itv realy means.
O.K.?
Now I asked in .0 if in the future DECmcc can do this translation (octet to
explanation string) for us. As you must be agree with me it is realy not
confortable always first to search the octet in the MIB file, to interprete it.
For example: In the CHIPCOM Priv MIB is
chipModClass03PortStatus 00 ---> means O.K
|
2306.3 | Mor INfos | ZUR01::SCHNEIDERR | | Tue Feb 11 1992 03:44 | 16 |
| Sorry Rahul,
Something went wrong and my reply started, befor I was finisched with wrinting.
O.K again the example:
chipcomModClass03PortStatus 00 ---> means "O.K."
chipcomModClass03PortStatus 01 ---> means "Link Failure"
If you look to the object, you get "00" or "01" or other oktets, whatever is possible.
Now we want to see direct the "O.K." or the "Link Failure".
Thanks Roland
|
2306.4 | Good idea, but not for V1.2... | TOOK::MINTZ | Erik Mintz, DECmcc Development, dtn 226-5033 | Tue Feb 11 1992 07:03 | 10 |
| That would be a handy feature.
Currently, each private extension has its own set of definitions.
We have no way of knowing which values correspond to which explanation strings.
Because the current MIB specification format does not include
a mapping from enumerated codes to explanation strings, we would
have to provide some sort of language to define the mappings.
-- Erik
|
2306.5 | | YAHEY::BOSE | | Tue Feb 11 1992 10:00 | 33 |
|
Actually, you could display text instead of the octet, but that
would mean fooling around with the mib defn. You could redeclare the
objects as ASN.1 enumerations, as follows.
chipModClass03PortStatus OBJECT-TYPE
SYNTAX INTEGER {
OK (0),
LINK_FAILURE (1),
PARTITION_JABBER (2),
REMOTE_LINK_FAILURE (3),
REMOTE_JABBER (4),
INVALID_DATA (5),
LOW_POWER (6),
FIFO_OVERRUN (7),
FIFO_UNDERRUN (8),
FATAL_ERROR (9),
PARTITION (10),
MJLP (11)
}
ACCESS read-only
STATUS mandatory
.
.
.
This would cause a corresponding enumeration to be created in the
MSL and the text would be displayed by MCC, instead of the actual
value.
Rahul.
|
2306.6 | Yea, but .... | ZUR01::SCHNEIDERR | | Wed Feb 12 1992 03:12 | 8 |
| Rahul,
Yes, I thought about this solution, too. But it,s realy not so sympatic to me.
If there would be something like a translationfile, or what ever, we hadn't
edit arround in the MIB defn files.
Thanks for your input Roland
|
2306.7 | It's the vendors problem. | SUBWAY::REILLY | Mike Reilly - New York Bank District | Wed Feb 12 1992 11:47 | 17 |
| Contact the vendor who produced the mib definition. Their mib
definition should have the enumerate values included. Until
customers start demanding quality snmp agents and quality
documentation these items will be neglected. Take the time to
contact the engineering manager at the company which produced this
mib and express your concern. I have personally delayed the order
of 40 multi-protocol routers because the vendor did not have a
valid ASN.1 definition for their mib. We received the the mib
definiton via express mail very soon after this.
When my current customer orders equipment which claims to be SNMP
manageable we demand that the vendor send a floppy with the current mib
definition with the product. If other customers asked for the same
treatment then many of these mib problems would be eliminated.
- Mike
|
2306.8 | Absolutely! | TOOK::MCPHERSON | Scientific progress goes 'Boink!' | Wed Feb 12 1992 12:37 | 5 |
| re .7
Bravo, Mike!
I agree 100%.
/doug
|
2306.9 | Yes, thanks | ZUR01::SCHNEIDERR | | Thu Feb 13 1992 08:01 | 7 |
| Mike,
Yes, your right, the vendor has to do some work, too.
Good input, thanks !!!!
Roland
|
2306.10 | yeah | TOOK::MATTHEWS | | Tue Feb 18 1992 15:22 | 6 |
| Mike, I fully support your comment. I have been giving the sme
advice to customers. Further, we need a "good housekeeping seal
of approval" for SNMP manageable objects to make sure they provide
good solutions.
wally
|