Link to home
Start Free TrialLog in
Avatar of macwalker1
macwalker1Flag for United States of America

asked on

Configuring APC Powerchute for Server Shutdown

Hi,

I have four APC 1500 units that I run their free Powewrchute on.  I am a bit confused as to what each of the configuration settings really mean.  

I want the UPS to keep the servers up as long as possible and then shut down.  However the way I currently have it configured it, I have it set to shut down immediately in the event of a power failure.

I have asked that the UPS units do the calibration but that appears to fail each time.

Any suggestions as to basic configuration?  

ON one in particular it reads that:
Run time is 64 mins.  So obviously I would like to take advantage of as much of that time as possible.

Here are the current parameters set on this particular machine:
Time Required For Shutdown Sequence  10 Minutes
 
  Low Battery Signal  120 Seconds
 
  Command File Enabled  No
 
  OS Shutdown Delay  1:30 Minutes:Seconds
 
  OS Shutdown Type  Shutdown
 
  Enable OS Reboot  Yes
 
  UPS Turn On Battery Capacity  30 %
 
  UPS Turn On Delay  0 Minutes
 
  Enable SMTP E-mail  Yes
 
  Enable Notification  No
 
  Enable Logging  Yes
 
I'm curious that maybe the "Shutdown Delay" may be too quick?  
Thanks
-Mac
ASKER CERTIFIED SOLUTION
Avatar of wwakefield
wwakefield
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of macwalker1

ASKER

Thanks wakefield.  On that setting, my maximum is about 10 mins.  I'm not making use of command line feature.  Just want the UPS to keep the machine up as long as possible.  Being on the uber safe side I had it set to shutdown if there were a powerfailure, period.  We clearly have a pwr issue as the pwr was off for only 6 seconds over the weekend according the log file in Powerchute.

Also, I think I have more than one server connected to one of these units and it appears that those machines came right back...so I'm still a bit confused and sorting out some things.  

Any ideas about optimal configuration here?  Or maybe even a better management of this?
wakefield:
After some more digging on one of my machines, it is configured to stay on as long as possible already.  I first thought I had configured to shutdown immediately but that was on another machine.

According to the logs the power was lost at 12:06:33 and restored 12:06:38,  so literally 5 seconds.  That being said, that machine shut down.  Any other ideas as to why this might have happened?
Optimal...   Just stating the obvious.   Whatever fits your needs.   In some of my overseas environments I set them to stay on 5 minutes since power is know to go off for 2 minutes then come back on.   If it is longer that 5 it will normaly be down for quite a while.

The setting/time will be unique per UPS since the batteries in each unit will be different based on where they are in their life cycle.   I like replacing them anually with a decent grade of battery.

If you are all onthe same network, you should/could run them all from the same single console session.   Could not tell if you were or not.

Power issue...    Have a look at the power quality in the charts on your UPS and turn that over to an electrical engineer or building management for review.   Crap power will also wear your batteries too.

Am I answering the right questions?

wakefield.

You are helping a lot.  They are all on the same network.  I have four APC 1500's.  My core machine is a SBS 2003, it is running the  APC Powerchute Business Edition Basic and is configured to stay on as long as possible and shows a "runtime" of 31 minutes.  The other "important" machine shows 64 mins remaining and it in fact shut down as well.  

Looking over logs and other info in the interface of each box now.  Just can't figure out why these two machines shut down completely while others came right back up.  Not to mention the fact it was literally only 5 seconds.  

I've also determined that the two most core machines we are talking about here lost power 3 minutes apart according to the log files.  But both for the same amount of time.

That give you anything to think about?
NOTE:  Make sure the server that did seem to work on the UPS properly is plugged into the proper plug on the UPS.  In other words not plugged into an UNSWITCHED outlet.
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
I'll check the "unswitched" option there. As for the CMOS...I don't know.

The machines that were "up" only required a CRL+ALT+DEL, but were "up".  It's these other two that are puzziling me.  It appears the one, not the SBS, did what I had it set to do, which was to shut down, but it's not the only box connected to this UPS and those seemed to have behaved as one would expect in such a quick interruption.  

To my knowledge "F1" hasn't been an issue.  

wakefield.

I just checked my connections and came up with this:

UPS 1
  SBS, FW and HP Switch are the only things connected to that UPS.  It says that the estimated run time is 31 mins.  This is one of the two that shut down.

UPS2
  Tape library and my NAS01 are the only things connected here.  The NAS01 is the other machine that shut down.  It has 2 pwr supplies.  only ONE of them are connected to this UPS.

UPS3
  My web server and my Citrix server are connected here and they were just waiting for a CTRL+ALT+DEL but were "up"

UPS 4
  The OTHER power supply for NAS01 is connected here and both pwr supplies to my NAS02 are connected here.  

So NAS01 had each of its power supplies connected to separate UPS units, but shut down.

SBS has both of its power supplies connected to one UPS and shut down.  

Does that give you any insight?  It seems I see something here but being a Monday I am a little slow.

Thanks
-Mac
So if it did SHUTDOWN and not just crash due to know power you can assume it is conencted to the UPS and the UPS inititated the shutdown.   You could always delete the console entry and readd it just to make sure it has the correct settings.
Both of these servers shut down.  I was called, was able to log into other machines, unable to browse to these.

Ended up trying to walk a user through powering up, NAS01 was no issue, but the SBS he said "is on but it's not" so I ultimately came in and booted the SBS.  I show no significant issues in the event viewers for either machine.  

Each ones log file in APC states "shutdown sequence started" immediately after the pwr issue is logged on each.
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Thanks.  I was about at that point as well.  Very strange.  Just looking at load on UPS 3.  Only two very lightweight servers on that one, load is at 43% and says run time is only 6 minutes.
Another option is to test it and watch the behaviour....

Pull the main plug and watch the servers come down after the battery drains.    Obviously probably a scheduled down time thing, but it would give you a chance to watch the behavior and see the POST when they come up.
wakefield.

I talked to APC and updated the agents and configured based on their recommendations.  They said that if in any of the options it was selected to "shut down" that even if power was restored as quickly as it had been in my case, that the process of shutting down would complete.  That it wouldn't interrupt that process.  So althought I had it set as I wanted, which was to shutdown, I guess I incorrectly assumed that if power was restored it would interrupt that process.  Foolish on my part I guess.

Thanks for all your help!

I'll award and close on this one.
Awesome...     Glad you got it!
My Solution for this exact problem that I was having with two NEW APC UPSs, one for each Server, where shutting down within seconds of Power failure no matter what I had put in the Shutdown configuration for delays. Was to not set the Shutdown option enabled for Power Failure under Event Actions,just notify and broadcast to all User and then under the Time on Battery Threshold Exceeded -decided 15min would be enough time on battery - then Shutdown - then under the Shutdown Sequence configuration - setup the delays to 3min OS Shutdown Delay, Time required for Shutdown Sequence set to  3mins.. this way... when on Battery -Powerchute waits for the 15min Threshold to be exceed (though you can set what you want), then starts the Shutdown Sequence- Then it waits 3mins for the shutdown to be completed and then powers down the UPS itself to save the Battery. When I plug the UPS back in.. assuming the BIOS is setup to enable Automatic Power On-(I had to go and enable on one of our servers) the Servers come back up. Hope this might help someone as I had to figure this out for myself. APC was still messing with the Shutdown Configuartion settings to make it wait.. with NO Luck.. when having the Event Action for Power Failure set to shutdown.. this was quite confusing but this is working exactly as I need it to. I think this what Macwalker1 the author was looking to solve also.