3300 Requires Hard Reboot on Power Failure

[kandekk]

We have a Mitel 3300 CX controller (circa 2005) on release 9.0.2.18 connected to a UPS.

When the power fails, and the outage is long enough that the UPS battery is out of juice, we need to do another hard reboot for the 3300 to properly boot up.

So, power comes back, the 3300 tries to boot up, but after several minutes the 3 alarm LEDs are still flashing red and the 3300 is stuck in this state until we do a hard reboot (pull the power cord). I've waited over an hour but it never leaves this state. This is reproducible as it happens every time there is power loss to the 3300.

Since this requires manual intervention, this is not an ideal scenario.

Ideally, the 3300 would work with the UPS similar to a Linux/Windows host and power down gracefully before the UPS battery drains, and then power up once power has been restored for N minutes. I've heard this can't be done with the 3300.

Is this normal behavior for the controller to require a hard reboot after power is restored? My VAR seems to think so.

Are there any power management tie ins with a UPS and Mitel's controllers (3300)?

Thanks for any info provided.
Chris

 

[IrwinMFletcher]

Do you have any console logs of it trying to start?

 

[tlpeter]

Any power outage is bad !!!
Before the power is pulled of you need to give the system the command "shutdown"
If you do not do that the harddisk could be damaged or the database corrupt !!!
Of course you can not control a power outage but you can control the UPS
If there are frequent power problems then talk to the power supplier or and take a bigger UPS
Ofcourse the CX should power up normaly but like i said pulling the plug is a bad thing to do with the "shutdown" command

ACS - Implement IP Office
ACA - Implement IP Telephony -- ACA - Design IP Telephony
ACA - Voice Services Management
______________
Women and cats can do as they please and men and dogs should relax and get used to the idea!

 

[kandekk]

Sadly I do not. I was in a rush to get the 3300 back online.
I guess I can reproduce the problem and get the console logs during bootup, but then I run the risk of breaking something (hard power resets can't be good for the controller and hard drive).

This has happened with previous releases we used to have (5.3), and with a different hard drive.

Am I safe in assuming this is not normal behavior of the older 3300's? I hope so.

Chris

 

[IrwinMFletcher]

No, it's not normal in the sense that you should be able to cold reboot a box without issue. There were some harddrive corruption issues, but I haven't heard of any in sometime. That being said, a backup is the sensible plan.

There were issue with some dsp modules that would get stuck (wouldn't come out of reset without a cold boot) but even on affected modules they were in the odds of 1:100000 of it happening so for it to consistently happen it would be odd. Even still, the behaviour was changed so that the system will come up without dsp modules (you can still make IP-IP calls).

If you can reproduce it the logs would help, other than that it's just guess work. You can either try to figure out why, or try to protect from it happening again.

 

[kandekk]

Ok, I will schedule an off-peak time to replicate and look through the console logs.

Preventing this from happening again would be ideal. If only there was some way for the UPS to notify the 3300 that it should gracefully shut down and then wake it back up once power is restored and stable. This provides a clean shutdown and I suspect that the 3300 would boot up properly after power is restored.

APC offers agents that run on Linux/Windows hosts to tell the host to power down/power up, but I didn't see anything relating to Mitel controllers (3300).

 

[drichter12]

Something to consider. Your UPS should provide for "delay on restore" type functionality which would prevent the UPS from restoring power to loads until a set delay period to ensure that utilities have completely restored. In other words, when power comes back on, the UPS should wait a set number of seconds before sending power to the load to help reduce the chance of power "bouncing" your equipment.



Dale

 

[kandekk]

Good point, and all our UPS's are set to wait a certain amount of time before restoring power to connected devices.

I think my path forward is:

1) Resolve the bootup issue on a hard shutdown by replicating and looking at the console logs. I'll bring my VAR in for this with a spare HDD just in case.

2) Develop/automate something to gracefully shutdown the 3300 in the event of a prolonged power failure. I still worry that the next time we lose power, the 3300 will go down hard and the hard drive breaks which means the 3300 is broken.

Thanks all for your responses.