Unsolved
This post is more than 5 years old
1 Message
0
3693
March 22nd, 2007 07:00
Down Vs. Unresponsive
We have had events on servers (discovered as hosts) where ICMP/SNMP are unavailable. For some servers, we have a 'DOWN' notification with an 'Unresponsive' event as part of the Impact. For other servers we have only 'Unresponsive' Events. And for others, we have 'Unresponsive' events that are IsRoot, then minutes later, a 'Down' event takes away the IsRoot flag for that server's event stream.
I'm trying to understand what type of diagnostics/tests that AM (6.2 sp1) does to escalate an event from Unresponsive to Down, and what is the difference between the 2 event definitions.
I understand that 'Unresponsive' is a symptom of a 'Down' event, but why would some events receive the 'Down' condition, and others not ?
Thanks
I'm trying to understand what type of diagnostics/tests that AM (6.2 sp1) does to escalate an event from Unresponsive to Down, and what is the difference between the 2 event definitions.
I understand that 'Unresponsive' is a symptom of a 'Down' event, but why would some events receive the 'Down' condition, and others not ?
Thanks
No Events found!
JDoherty2
2 Posts
0
May 18th, 2007 05:00
Unresponsive can be a symptom of Down, but a system can be unresponsive but not down.
DavidRedwine
2 Posts
0
October 24th, 2009 13:00
"Unresponsive means it doesn't respond to SNMP requests. Down means it doesn't respond to ICMP requests,
Unresponsive can be a symptom of Down, but a system can be unresponsive but not down."
I have this to add: (Dave R.)
Down indicates that a system has failed. For example, Router Down indicates that a router has failed. A system failure causes all ports or interfaces on the system and all objects accessed through the system to be unreachable.
The events used as symptoms to diagnose Down vary, depending on whether the system is connected to other systems by a physical or logical link. When a system is not connected to other systems, the event is system Unresponsive.
All customer networks should have contiguous network maps, if they do not then you need to address that. Either a Switch, Router, Firewall or other intermediate device is missing or you have not been provided the intermediate subnet to be used in the discovery.
When a system is connected to one or more systems, the events used as symptoms to diagnose system Down include:
◆ MightBeDown for the system itself and any connected systems.
◆ DownOrFlapping for any peer network adapter and any peer subinterfaces.
DownOrFlapping means that the device that hosts the interface has sent 3 or more link/up/down traps on that interface within a 5 minute time period.
SNMP-Agents will show "NotResponding" when they are unresponsive to SNMP queries.
DavidRedwine
2 Posts
0
October 26th, 2009 07:00
Down indicates that a system has failed. For example, Router Down indicates that a router has failed. A system failure causes all ports or interfaces on the system and all objects accessed through the system to be unreachable.
The events used as symptoms to diagnose Down vary, depending on whether the system is connected to other systems by a physical or logical link. When a system is not connected to other systems, the event (symptom) is system Unresponsive.
When a system is connected to one or more systems, the events used as symptoms to diagnose system Down include:
◆ MightBeDown for the system itself and any connected systems.
◆ DownOrFlapping for any peer network adapter and any peer subinterfaces.
Reference Page 61 (3-11) of the EMC Smarts IP Availability Manager User's guide.