PCNS 4.2 shutting down VMware ESXi hosts even though power is available

[Anonymous user]

I have a problem where after moving the PCNS VMA's (virtual machine appliance) to a new host, the managed servers continuously get shutdown (cleanly) even though power is not out.

I have 6 production servers running VMware ESXi 5.x (with licensed versions  of vsphere standard or essentials.)      They are divided into two clusters of 3 servers each.  Each cluster has two APCC smart UPS'  SU2200 with network management cards (APC 9630/9631) .  Each server has dual power supplies  in a redundant connection.    (i.e. each server plugs into two UPS's.)     I have a management server (also running ESXi 5.x) that is NOT in any cluster.  The management server has 4 VMA (virtual management appliance) VM's-   2 vsphere servers (to manage each cluster) and 2 VMA's with APCC PCNS 4.2 for ESXi to shutdown the clusters when the power goes out.

This weekend I moved the   VMA's from the management server to a replacement management server (running ESXi 5.5 )  VMA's kept the same MAC and IP addresses.   The replacement management server was in a third cluster.    But no VMA was managing in cluster in which itself was installed.  When I started the APC VMA's backup all 6 production servers did a graceful shutdown.       I brought them backup and they immediately shutdown again.   I moved the VMA's back to the original host, brought the production servers back up, and they immediately where shutdown again.    I  had to shutdown the APC VMA's to prevent this happening.

According to my notes, I enabled shutdown under the following events (ideally keeping the systems up for 10 minutes on battery.)

• on battery (with a delay option available, default 120 sec- change to 600 seconds and enabled

I had the following two options but cannot remember if I actually enabled them

• run time exceeded (with a delay option available, default 0 sec)
• run time below threshold (with a delay option available, default 0 sec)

I was worried it may cause a shutdown the power was still on but the server but there wasn't enough battery charge to run the systems for 10 min in case of power outage.      But it seems to make sense that I would want to shutdown the systems if I am on battery but don't have enough for 10 minutes.   The two options seem pretty similar  and not well defined in the documentation.   Can't easily check the VMA settings without starting them up.

I don't think I have had a power outage since I last set the APC settings so this is probably the first time I restarted the APC PCNS VMA's. 

Any idea what is going on ?

Thanks

[BillP · Accepted solution]

Hi,

Open the pcnsconfig.ini file and verify that event_RunTimeExceeded_enableShutdown = false. The issue you are describing sounds like the setting is set to true so when PCNS starts it see that runtime has been exceeded and begins a shutdown.

Description of Available runtime has been exceeded

For both conditions below, the “total shutdown time” includes the following durations:
• VM migration duration
• VM shutdown and startup duration
• vApp shutdown and startup duration
• vCenter Server VM shutdown duration
• Shutdown command file duration
• Built-in duration of 2 minutes (this consists of a 10 second OS shutdown duration and a 60 second OS shutdown
duration; rounded up)

This event occurs with either of the following conditions:

Condition 1.
When the total shutdown time required by PowerChute is greater than the Low Battery Duration minus two minutes
configured for the UPS. In the event of a low battery condition, PowerChute will not have enough time to complete
the shutdown sequence before the UPS powers off. For example, if the total shutdown time required is 3 minutes
and Low Battery Duration is 4 minutes, the Available Runtime has been Exceeded event will be triggered.

Resolution:

Increase the Low Battery Duration value on the NMC using Configure - Shutdown or decrease the
shutdown durations being used by PowerChute.

Condition 2.
When the shutdown duration configured for the UPS On Battery event plus the total shutdown time required by
PowerChute is greater than the Runtime Remaining on the UPS minus two minutes. This condition can be caused by
having too great a load on the UPS when the battery is fully charged.

Resolution:
1. Remove some equipment from the UPS to increase the available runtime.
2. Decrease the shutdown duration time for the UPS On Battery event.
3. Decrease the command file execution time using the Shutdown Settings screen.
This event is logged and event actions are carried out even if it occurs on a single UPS in a Redundant or Parallel
UPS configuration.