[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
| Title: | DECmcc user notes file. Does not replace IPMT. |
| Notice: | Use IPMT for problems. Newsletter location in note 6187 |
| Moderator: | TAEC::BEROUD |
|
| Created: | Mon Aug 21 1989 |
| Last Modified: | Wed Jun 04 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 6497 |
| Total number of notes: | 27359 |
5127.0. "Time problems with historian." by KETJE::PACCO (Gallia divisa est in partes tres) Fri May 28 1993 12:26
I have problems similar to the ones reported under note #4645,
but in a totally different context.
4 historian related problems:
1). The data is not retrieved for the time specified, from the
historical MIR, but is only retrieved from the runing network.
2). A show command with on historical data is executed "now".
3). A show command with multiple "start time" is not executed at all.
4). Show recording does not accept the same entity data types as the
normal "show command"
1). The data is not retrieved for the time specified, from the
historical MIR, but is only retrieved from the runing network.
======================================================================
From the iconic map I started recording on entity
"node rgkk01 hdlc link rrhc01 logical station rgkk01"
and I gave as partition "*"
From the fcl_pm I see:
MCC> sho rec node .gkk.m.rgkk01 hdlc link rrhc01 logical station
rgkk01,
in domain .mcc.gkk
PARTITION = *
Node bc:.gkk.m.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-16:41:24.698
Recording parameters for the partition "Identifiers" :
Recording state = ACTIVE,
State since = 1993-05-28-15:54:34.420,
Polling period = +0-00:10:00.000I-----,
Begin time = 1993-05-28-15:54:25.483,
End time = 2012-05-25-00:00:00.000,
Keep age = +45-00:00:00.000I-----,
Request time = 1993-05-28-15:54:25.483,
Requested by = "root",
Time of last successful poll = "1993-05-28-16:34:42.534",
Number of successful polls = 5,
Time of last failed poll = "NONE",
Last poll failure reason = "N/A",
Number of failed polls = 0,
Last historical expiration time = "1993-05-28-16:34:42.534",
Time of last expiration failure = "NONE",
Last expiration failure reason = "N/A",
Number of expiration failures = 0
MCC>
>>> Hence, the recording was started only for Identifiers.
2). A show command with on historical data is executed "now".
=============================================================
MCC> sho rec node .gkk.m.rgkk01 hdlc link rrhc01 logical station
rgkk01,
in domain .mcc.gkk
PARTITION = counters
Node bc:.gkk.m.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-17:01:13.664
Recording parameters for the partition "Counters" :
Recording state = ACTIVE,
State since = 1993-05-28-16:46:04.440,
Polling period = +0-00:15:00.000I-----,
Begin time = 1993-05-28-16:45:56.897,
End time = 2012-05-25-00:00:00.000,
Keep age = +45-00:00:00.000I-----,
Request time = 1993-05-28-16:45:56.897,
Requested by = "root",
Time of last successful poll = "1993-05-28-17:01:09.684",
Number of successful polls = 2,
Time of last failed poll = "NONE",
Last poll failure reason = "N/A",
Number of failed polls = 0,
Last historical expiration time = "1993-05-28-17:01:09.684",
Time of last expiration failure = "NONE",
Last expiration failure reason = "N/A",
Number of expiration failures = 0
MCC> sho node .gkk.m.rgkk01 hdlc link rrhc01 logical station rgkk01 all
cou, for start=(17:00)
Node bc:.gkk.m.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-17:01:58.250 Counters
Examination of attributes shows:
Creation Time = 1993-03-26-11:47:04.042
Data Octets Received = 262925057
Data Octets Sent = 462250210
Data PDUs Received = 3626638
>>> Data PDUs Sent = 3816322
FRMRs Received = 0
......
MCC> sho node .gkk.m.rgkk01 hdlc link rrhc01 logical station rgkk01 all
cou, for start=(17:00)
Node bc:.gkk.m.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-17:04:28.516 Counters
Examination of attributes shows:
Creation Time = 1993-03-26-11:47:04.042
Data Octets Received = 262930761
Data Octets Sent = 462257870
Data PDUs Received = 3626733
>>> Data PDUs Sent = 3816410
FRMRs Received = 0
......
>>> Both the "AT" time and the incrementing "Data PDUs Sent" is the
>>> proof that the "Show" command is executed "now" instead on past
data.
3). A show command with multiple "start time" is not executed at all.
======================================================================
MCC> sho node .gkk.m.rgkk01 hdlc link rrhc01 logical station rgkk01 all
cou, for start=(17:00)
Node bc:.gkk.m.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-17:04:28.516 Counters
Examination of attributes shows:
Creation Time = 1993-03-26-11:47:04.042
Data Octets Received = 262930761
Data Octets Sent = 462257870
Data PDUs Received = 3626733
Data PDUs Sent = 3816410
FRMRs Received = 0
......
MCC> sho node .gkk.m.rgkk01 hdlc link rrhc01 logical station rgkk01 all
cou, for start=(16:59,17:00)
Node bc:.gkk.m.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-17:07:50.411 Counters
Requested data not available for specified time
4). Show recording does not accept the same entity data types as the
normal "show command":
====================================================================
MCC> sho node rgkk01 hdlc link rrhc01 logical station rgkk01 all id
Node bc:.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-16:35:54.205 Identifiers
Examination of attributes shows:
Name = rgkk01
MCC> sho node .gkk.m.rgkk01 hdlc link rrhc01 logical station rgkk01 all
id
Node bc:.gkk.m.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-16:36:08.604 Identifiers
Examination of attributes shows:
Name = rgkk01
MCC> sho rec node .gkk.m.rgkk01 hdlc link rrhc01 logical station
rgkk01,
in domain .mcc.gkk
PARTITION = identifiers
Node bc:.gkk.m.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-16:36:32.815
Recording parameters for the partition "Identifiers" :
Recording state = ACTIVE,
State since = 1993-05-28-15:54:34.420,
Polling period = +0-00:10:00.000I-----,
Begin time = 1993-05-28-15:54:25.483,
End time = 2012-05-25-00:00:00.000,
Keep age = +45-00:00:00.000I-----,
Request time = 1993-05-28-15:54:25.483,
Requested by = "root",
Time of last successful poll = "1993-05-28-16:34:42.534",
Number of successful polls = 5,
Time of last failed poll = "NONE",
Last poll failure reason = "N/A",
Number of failed polls = 0,
Last historical expiration time = "1993-05-28-16:34:42.534",
Time of last expiration failure = "NONE",
Last expiration failure reason = "N/A",
Number of expiration failures = 0
MCC> sho rec node rgkk01 hdlc link rrhc01 logical station rgkk01, in
domain .mcc.gkk
PARTITION = identifiers
Node bc:.rgkk01 HDLC Link rrhc01 Logical Station rgkk01
AT 1993-05-28-16:36:51.030
No recording for the specified entity and partition
>>> Clearly, the "synonym" is interpreted as "DNS fullname"
Doesn't it looks strange ?
Dominique.
| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 5127.1 | | TOOK::SHMUYLOVICH | | Fri May 28 1993 18:23 | 53 |
|
> From the iconic map I started recording on entity
> "node rgkk01 hdlc link rrhc01 logical station rgkk01"
> and I gave as partition "*"
> .
> .
> .
> >>> Hence, the recording was started only for Identifiers.
What version of MCC are you running?
What messages did you receive from IMPM as a result of a Record command?
Did you try to create multiple recordings from FCL?
> 2). A show command with on historical data is executed "now".
> =============================================================
> .
> .
> >>> Both the "AT" time and the incrementing "Data PDUs Sent" is the
> >>> proof that the "Show" command is executed "now" instead on past
> data.
It happens because you forgot to specify In Domain qualifier. Please,
enter the Show command as follow:
MCC> sho node .gkk.m.rgkk01 hdlc link rrhc01 logical station rgkk01 all
cou, for start=(17:00) , in domain .mcc.gkk
--------------------
> 3). A show command with multiple "start time" is not executed at all.
> ======================================================================
See previous item.
> 4). Show recording does not accept the same entity data types as the
> normal "show command":
> ....
> >>> Clearly, the "synonym" is interpreted as "DNS fullname"
Let me briefly explain how the Historian support multiple identifiers
for a target entity.
When the Record command is specified among other actions it creates
an entry for the specified target entity in its private database. This
entry is used by any Historian's command to locate a corresponding
recording request. In order to accomplish this task the Historian uses
Directory and Show All Ident commands to receive all known identifiers.
The bottom line is that the Historian does not interpret identifiers
for a target entities but uses only what the corresponding AM is returning.
Hope this helps, Sam
|
| 5127.2 | OK and NOT OK. | KETJE::PACCO | Gallia divisa est in partes tres | Tue Jun 01 1993 12:33 | 80 |
| SAM,
>> From the iconic map I started recording on entity
>> "node rgkk01 hdlc link rrhc01 logical station rgkk01"
>> and I gave as partition "*"
>> .
>> .
>> .
>> >>> Hence, the recording was started only for Identifiers.
> What version of MCC are you running?
> What messages did you receive from IMPM as a result of a Record command?
> Did you try to create multiple recordings from FCL?
DECmcc V1.3/ULTRIX
Recording of "Identifiers" partition started
Individually requested recordings for "Ident", "Chars", ...
records ultimately simultaneously. It's only the wildcard
partition which does not translate in "all" partitions.
>> 2). A show command with on historical data is executed "now".
>> =============================================================
>> .
>> .
>> >>> Both the "AT" time and the incrementing "Data PDUs Sent" is the
>> >>> proof that the "Show" command is executed "now" instead on past
>> data.
> It happens because you forgot to specify In Domain qualifier. Please,
> enter the Show command as follow:
> MCC> sho node .gkk.m.rgkk01 hdlc link rrhc01 logical station rgkk01 all
> cou, for start=(17:00) , in domain .mcc.gkk
> --------------------
Right, because the FCL syntax is heavy (but clear and logical)
at the end: "mea culpa, errare humanum est".
But is is confusing to accept the "show" command with a for clause,
which is ignored (no error reported) when the "in domain" is not
specified, where as when "in domain" is specified it is used.
> 3). A show command with multiple "start time" is not executed at all.
> ======================================================================
See previous item.
>> 4). Show recording does not accept the same entity data types as the
>> normal "show command":
>> ....
>> >>> Clearly, the "synonym" is interpreted as "DNS fullname"
> Let me briefly explain how the Historian support multiple identifiers
> for a target entity.
> When the Record command is specified among other actions it creates
> an entry for the specified target entity in its private database. This
> entry is used by any Historian's command to locate a corresponding
> recording request. In order to accomplish this task the Historian uses
> Directory and Show All Ident commands to receive all known identifiers.
> The bottom line is that the Historian does not interpret identifiers
> for a target entities but uses only what the corresponding AM is returning.
The problem for the end user (which includes ALL DECmcc managers)
is that the "synonym" is ACCEPTED as valid entity value in all "show"
calls but NOT ACCEPTED in export, record calls.
It can be EXPLAINED because the "Synonym" is not an "Identifier", but
it is once again confusing. DECmcc is already soo complex, that these
little things only makes it use more difficult.
I hope that these and other comments are being taken care of for
new DECmcc (DME) versions.
Thanks,
Dominique.
|
| 5127.3 | | TOOK::SHMUYLOVICH | | Wed Jun 02 1993 18:01 | 35 |
|
> ...It's only the wildcard partition which does not translate in
> "all" partitions.
Please, wait until Friday afternoon. Sorry but I'm not able to check
it earlier.
> But is is confusing to accept the "show" command with a for clause,
> which is ignored (no error reported) when the "in domain" is not
> specified, where as when "in domain" is specified it is used.
When the "in Domain" qualifier is not specified the default domain is
used. The name of this domain is "DEFAULT". This can be overridden
by the logical (environment variable) MCC_DEFAULT_DOMAIN_NAME.
> The problem for the end user (which includes ALL DECmcc managers)
> is that the "synonym" is ACCEPTED as valid entity value in all "show"
> calls but NOT ACCEPTED in export, record calls.
>
> It can be EXPLAINED because the "Synonym" is not an "Identifier", but
> it is once again confusing. DECmcc is already soo complex, that these
> little things only makes it use more difficult.
From mine point of view the Historian as a generic FM has no
knowledge about which synonym is acceptable which is not. It's
It's responsibility of AM to advertise how the corresponding entity
can be called.
But I do agree that this is very confusing and should be taken care
in "new DECmcc (DME) versions".
Thank you for pointing to this problem.
SAm
|
| 5127.4 | The DECnetV AM is not consistent on this point ! | KETJE::PACCO | Gallia divisa est in partes tres | Fri Jun 04 1993 13:13 | 8 |
| Sam,
For the last point ...
I agree with you that there is no problem with the historian. It is
rather the DECnet V AM which has to advertise its synonym as
identifier, or not accept the synonym as valid entity value.
Dominique.
|
| 5127.5 | What the AM uses synonyms for | TOOK::PURRETTA | | Fri Jun 04 1993 14:26 | 23 |
| > or not accept the synonym as valid entity value
The AM doesn't do windows :^)
The DNA5 AM *always* sees the request come in as a fullname, even if you
specified a synonym on the command line. The PM generates the fullname
before calling us. And, we *always* return a fullname. That is our Identifier.
The DNA5 AM needs/uses the synonym for one thing. Connections.
When you register a phase V node, you are required to provide a synonym arg.
During the register process, the registration FM hands this synonym to the
DNA5 AM, and we store it in DNS under the global entity. Then when a request
comes in later on and we can't make a connection to the node via other methods,
we go to DNS and grab the synonym. We give this synonym to DECnet (be it
phase IV or phase V; We do all our connections in phase IV compatibility
mode.) and let it resolve how to get to the node. That's what/how the *AM*
uses synonyms.
I believe if you consistantly use the fullname when dealing with phase V
entities you won't have a problem.
John
|
| 5127.6 | Multiple recording from Iconic Map using "*" in partition | TOOK::SHMUYLOVICH | | Tue Jun 08 1993 10:05 | 13 |
|
Sorry for such a slow response about a wildcard in the "partition"
argument in the Record command entered from Iconic Map.
It seems to me that your problem occurs because you did not
use the button "MORE" which appears when the first response
( with "Identifiers" partition) is returned. In order to record
another partition you should click on this button. The next
response will be with another partition and so on. The button
"MORE" disappears when recording for all partitions are created.
Sam
|
| 5127.7 | Today I learned again something from DECmcc. | KETJE::PACCO | Gallia divisa est in partes tres | Wed Jun 09 1993 14:11 | 12 |
| Thanks Sam,
This is again something I forgot too look at. You are right, I didn't
hit the "MORE" button. You really have to work with DECmcc day and
night to arrive at the right experience. This is very hard.
Specially to coop with new releases, with new nuts and bolds.
Thanks,
Dominique.
|