| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 5214.1 | | MOLAR::YAHEY::BOSE | | Thu Jun 17 1993 11:13 | 8 |
|
This is the first time I've heard this. Can you fill me in on the
Entity instance, event type, and time stamps on each of these
multiple events? Do you have any alarm rules written against
reachability events? (Reachability events are logged automatically
since they are notification events, and do not need alarm rules).
Rahul.
|
| 5214.2 | Another problem | ANTIK::WESTERBERG | Stefan Westerberg DS Stockholm | Fri Jun 18 1993 06:16 | 13 |
|
Hi, I have alos seen a problem with the IP reachability events.
If a reachability state change is detected by the IP Poller
on an SNMP entity that is registrated in both domain A and B.
The event is delivered to domain B even when we don't have
the ip poller checking that domain. The A and B domain isn't
in the same domain tree.
The operator of domain B don't want events from domain A.
Today we use the filter option as a work around to suppress all
IP reachability events in domain B.
/Stefan
|
| 5214.3 | | MOLAR::YAHEY::BOSE | | Fri Jun 18 1993 16:00 | 9 |
|
The IP Poller does an mcc_event_put to the event pool when the
reachability status changes. Since this is a notification event,
the Notification FM picks up this event and notifies the domains
where the entity reside. So if you have the entity in question
residing in two different domains, both the domains will be notified
in the current scheme of things.
Rahul.
|
| 5214.4 | | BIGUN::MAYNE | `AXP!': Bill the Cat | Mon Jun 21 1993 05:05 | 8 |
| The nodes each live in one domain, but there are three domains with reachability
polling turned on (plus another where it isn't).
The time stamps on each of three alarms for any one event are the same. There
are no reachability rules. Event types are IP reachability up/down, instances
are SNMP nodes.
PJDM
|
| 5214.5 | Same thing seen when polling ip nodes in single domain | CUJO::HILL | Dan Hill-Net.Mgt.-Customer Resident | Wed Jun 23 1993 02:04 | 21 |
| There is definitely something wrong, but the problem is more aggrevated
when ICMP retries is 2 or less and the ICMP timeout is 10 (at least in
my case).
It appears that the Notification FM is displaying configuration events
as a result of the poller receiving and processing 2 (or more) ICMP
echo replies. I have seen this occur for a node that is not even down.
I'm monitoring 228 IP nodes in one domain using the following IP Poller
settings:
Period = 90 sec
Retries = 5
Timeout = 5 sec
If I get any lower than 90 seconds, I start getting nodes incorrectly
reported as "not responding", but only a couple.
The "double" configuration events occur more frequently if I set
retries to 2 or less.
-Dan
|
| 5214.6 | | MOLAR::YAHEY::BOSE | | Wed Jul 07 1993 12:35 | 12 |
|
The IP Poller sends out 50 ICMP echo requests at a time and waits
for the echo replies till the timeout period expires. It seems that
the receive buffer in the current poller is too small, thus causing
some of the echo replies to be dropped. This is why you might see
nodes show up as unreachable, although they actually aren't. If you
have a high enough number of retries, the lost packets get resent
and the poller appears to work correctly. I have increased the size
of receive buffers so that this behaviour does not occur any further.
Rahul.
|
| 5214.7 | Please supply fix to the field (and trade shows) | TOOK::R_SPENCE | Nets don't fail me now... | Wed Jul 07 1993 15:52 | 4 |
| Will there be a field update kit? Field sites need the fix asap.
s/rob
|
| 5214.8 | | MOLAR::YAHEY::BOSE | | Wed Jul 07 1993 16:53 | 5 |
|
Please send me mail if you need a copy of the executable with the
fix in it.
Rahul.
|
| 5214.9 | | BIGUN::MAYNE | `AXP!': Bill the Cat | Tue Jul 27 1993 02:55 | 3 |
| But does this fix the multiple occurences, as opposed to missing occurences?
PJDM
|
| 5214.10 | | MOLAR::YAHEY::BOSE | | Tue Jul 27 1993 10:50 | 7 |
|
I still haven't been able to reproduce the multiple notification
problem. Are you sure you have only one notify request enabled?
Please send me mail with all the details so that I can try and
reproduce the problem on my system.
Rahul.
|
| 5214.11 | Is patch mentioned in .8 available | MADMXX::WOESTEMEYER | Why??...Why not!!! | Wed Nov 02 1994 10:56 | 18 |
| I have a customer who is reporting the same multiple notification from
the IP poller. I have verified that no other alarms are executing. Attached
is his mcc_notification.log.
Steve Woestemeyer
dtn 592-4208
Alarm: critical SNMP TESTER IP Reachability Down 2-NOV-1994 09:02
:58.87 Domain GAIA_NS:.9-11_LINCOLN [14,1]
Alarm: critical SNMP TESTER IP Reachability Down 2-NOV-1994 09:02
:58.88 Domain GAIA_NS:.9-11_LINCOLN [9,2]
Alarm: critical SNMP TESTER IP Reachability Down 2-NOV-1994 09:02
:58.89 Domain GAIA_NS:.9-11_LINCOLN [17,3]
Alarm: clear SNMP TESTER IP Reachability Up 2-NOV-1994 09:02:59.52
Domain GAIA_NS:.9-11_LINCOLN [14,4]
Alarm: clear SNMP TESTER IP Reachability Up 2-NOV-1994 09:02:59.53
Domain GAIA_NS:.9-11_LINCOLN [9,5]
Alarm: clear SNMP TESTER IP Reachability Up 2-NOV-1994 09:02:59.54
|