Hi
I get this event every 5 min. from my vCenter Server but only for some off mine 18 host.??
When restart vCenter Server, the event disapear for wile??
Thanks
Henrik Froulund Hansen
-
Target: esx07.int.unicon.dk
Stateless event alarm
Alarm Definition:
Event details:
Error detected on esx07.int.unicon.dk in Brondby: Agent can't send heartbeats.msg size: 612, sendto() returned: Operation not permitted
-
The errors are always on the same hosts?
All 18 hosts are in a same cluster?
Or different clusters in the same datacenter?
Andre
It seems to be a problem when I reboot any off the host.
And the messages disapear for all the serveres, if I restart the vCenter service!!
There is no alarm in the client!
/Henrik
You have a single cluster?
Which version of ESX and VC are you using?
Andre
One cluster with 5 host + 13 host with out cluster
It's ESX4.0 and VC 4.0.0 build 162856 :o))
This is from the latest test:
-
Target: esx01.int.unicon.dk
Stateless event alarm
Alarm Definition:
Event details:
Host esx01.int.unicon.dk in Brondby is not responding
-
And after a restart off VC-server
-
Target: Datacenters
Stateless event alarm
Alarm Definition:
Event details:
VCSERVER02.int.unicon.dk (vCenter) status changed from red to green
-
But there is now alarm inside the VC-client???
Regards/
Henrik
Hey froulund --
This happens to me too and is VERY annoying (alerts are forwarded to my team's mobile devices). I turned on the alarm for "Host Error", and I believe it's a repeating alarm by default. (Obviously if you set it not to repeat it will be MUCH less annoying.) As for the root cause, I don't know for sure, but here's what I've been able to determine:
1) It happens at roughly the same time of day when it does happen.
2) It seems to be related to a scheduled task with VMware Update Manager scanning hosts (6:00AM every day for me), and when VUM scans the hosts for updates and patches, it appears that the vc.Integrity (agent?) modifies the host firewall in real time so it can perform the scan.
3) I can reproduce the error in my environment on random hosts in any cluster by running the scheduled task manually.
4) It seems that this is a transient error that resolves itself shortly after the alarm condition is created once the host firewall is "re-adjusted".
WORKAROUND: My solution that works thus far is to modify the "Host error" alarm definition by adding an advanced host error trigger condition for "Full message" that is not equal to the string "Agent can't send heartbeats.msg size". Voila. Annoying alarm pages / emails averted but still logged, and transient condition passes without freaking everybody out or annoying the crap out of everybody...
If this is helpful or "correct", let me know! :smileycool:
UPDATE: The error seems to come most often from the vSphere host that my vCenter VM is running on. I vmotioned it to another host, re-ran the update task, and the heartbeats.msg host error now occurs on the new host running my vCenter VM. Still got no "alarms" or annoying pages tho... :smileygrin:
Regards,
Pete Kowalsky - VCP3, CISSP, CCNP, CCSP, ADHD, blah blah blah
(Points for helpful / correct are appreciated...)
Hi
After I deletet all in "host error" re instertet them i don't have any "fake" alarms.
So right now I am happy
Thanks to all and have a nice sommer!
/Henrik
Hello,
also in my case the fake alarm stopped after recreating Host Error from triggered alarm.
Marco.
Why not, as a workaround, restart the failing service after the update proces has finished?
Would that not be more elegant solution?