Enterprise Manager reporting lost connectivity to nodes 1

[fonejack]

Has anyone else run across a problem of enterprise Manager reporting "Lost connectivity to Node" when connectivity has not been lost?? We continually run into this problem, selecting update now does not fix it. To test I setup a constant ping to the node from the eMan server and the pings never fail. The constant false positives make the alarm reporting useless.

Any ideas or has anyone else run into this problem? Thanks!

 

[IrwinMFletcher]

I have no experience with EntMgr, but on the surface remember that, 'ping' does not equate to connectivity between applications. Ping just lets you know that you can reach one interface from the other.

How long after it's connected does it stay like that before it loses the node? Anything else running that may block the ports? Firewalls, etc?

 

[fonejack]

I believe EM uses ping to establish whether or not nodes are reachable and then outputs the node connectivity alarm if the ping fails. The test is performed every 5 minutes and clears the alarms randomly.

 

[MitelInMyBlood]

We've had the problem and it quickly became such a nuisance we finally turned off status polling. Alarms coming and going throughout the day and night for no apparent reason. Our network is not going up and down. Our ntwk monitoring systems would tell us if it was.

In my opinion..... EMgr still has a few warts on it. It works well for what we use it for (alarm reporting of valid system alarms) but whether a site chooses to respond to an ad-hoc status poll or not has had to be removed from an item of concern for us because (again in my opinion) status polling in Emgr is unstable/unreliable.

Escalation policies in Emgr are also unreliable, at least as they apply to the SX2K redundant sites. Product Support says they can't dupe it and we've simply grown weary of opening up ticket after ticket on it. At one time we had an area TAM who would come on site, witness a nagging problem and see it through to resolution. That was the old, pre-Intertel Mitel. In another couple years all my 2K Redun sites will be gone and it will thereafter be a non-issue. For now, like status polling, we've decided it's an acceptable risk. 2K Redundant sites (on stable power) have problems so dang seldom that it's not worth taking blood pressure meds to fight over. After 42% layoff there's no one here who even has time to care.



Original MUG/NAMU Charter Member

 

[fonejack]

Thanks for the info, glad to know it’s not just me. I’ll disable polling and move on

 

[MitelInMyBlood]

That's what I did. If a box out there dies you're going to get alarms from other interconnected machines anyway and if it's standalone and goes belly-up you're still going to get a call from someone. If you're a VAR I wouldn't try marketing this to clients offering to do central site monitoring of their sites for them, but otherwise it's very handy for being able to see the entire network (or any portion) at a glance and click on a site to get onto it quick without having to remember all the IP addresses. It has it's value, but also has it's issues.

Original MUG/NAMU Charter Member

 

[fonejack]

One more question; now that I have turned off status polling the constant Lost connectivity node alarms have stopped but now the regular alarms have stopped updating also. Is there any way to get the real alarms to report in real time while polling is turned off?

 

[Lundah]

Instead of just turning off the status updates, you can also tweak the SNMP polling settings a bit. In the AutoDiscovery Configuration menu, select the "Protocol" tab, click "SNMP", and then the "Properties" button. Change the SNMP Timeout field to 5 seconds (max setting), and the SNMP Retries field to 3 (again, the max setting). This should cut down on the bogus alarms a LOT.

 

[kwbMitel]

Lundah, Good Info, thanks.

What are the default values, I don't have a system immediately avail.

Fonejack:
Eman has an Alarm Client that can be downloaded and installed on the PC. This would get the alarm status in realtime directly from the controllers (I think, not sure)



*******************************************************
Occam's Razor - All things being equal, the simplest solution is the right one.

 

[mclise]

Make sure the SNMP is turned on on the PBX and the community strings on the PBX and EM are the same.

I was having the same problem and that was the fix.

 

[kwbMitel]

Lundah,

I just adjusted the Auto-Discovery values for a site that was experiencing this issue.

For the record, the Default Wait time is 2 seconds and the retries is 0(zero)

I've given you a star as this has been a pain point for me for quite some time. (I'm on a email distribution for alarms for this site)

*******************************************************
Occam's Razor - All things being equal, the simplest solution is the right one.