UPS communication lost error - vMA on ESX 5.1

[Anonymous user]

Hello everyone,

I just joined this forum specifically to address this problem.  We have an HP Proliant ML350P Gen 8 with an ESX 5.1 installed on it.  We will eventually be converting a Windows Server 2003 VM on it and we need to add a graceful shutdown option to the server.  We bought a SmartUPS 1500 to manage the power for the server.  I then followed the instructions found here in the link below to install vMA and set up PowerChute Business Edition:

http://www2.schneider-electric.com/resources/sites/SCHNEIDER_ELECTRIC/content/live/FAQS/165000/FA165...

Obviously since I'm installing on a linux machine (vMA) running SUSe, I have to use the RJ-45 to Serial cable. The installation goes well and once it's done I'm able to manage from the web browser for a certain period of time.  The problem is when I come in the next day I can no longer view the UPS status in the PCBE web browser.  When I log in, a message says "UPS Communication lost".  I have already tried the following steps:

1) Verify the UPS-to-server cable is correct.  I am using the suggested 940-0625 cable.  I have tried re-seating it and even changing to a backup cable of the same type (yes, it is also an APC cable).  None of these restore the communication.

2) Restart the services on the vMA.  I do this with two methods:

                  sudo /etc/rc.d/init.d/PBEAgent stop, sudo /etc/rc.d/init.d/PBEAgent start

                  ./opt/APC/PowerChuteBusinessEdition/Agent/bin/ startup stop /opt/APC/PowerChuteBusinessEdition/Agent/bin/ startup start

     The reset is working since I lose access to the PCBE web session but when I log back in after the service is up the UPS communication is still lost.

3) Reset machines.  I tried resetting the vMA, and even the server.  When they come back online there is no change in the status.

4) The ONLY option that works so far is to completely power everything off, turn off the UPS, uplug and replug and then power everything back on.  Once I do this, everything returns to normal and I have access to all PCBE's functions.  This is not an optimal solution since it once again loses communication the following day.

I have been troubleshooting this for over a week now and it is very frustrating.  I even tried reinstalling the vMA on a completely new UPS on a completely different Dell server with the same result.  Obviously there is something wrong with the configuration so I called Schneider to open a ticket.  We walked through the same processes above with the same result.  One suggested solution was to use a USB cable, but the linux install does not give that option.  It only allows a DB9 or RJ-45.  Since our UPS only comes with an RJ-45 and a USB type B, we have no choice but to use RJ-45 to Serial (unless there is some sort of RJ-45 to USB cable that APC supports).

Does anyone else have any ideas about how to get this thing working?  The only difference between the instructions above and the way I did things was to use thin provisioning instead of thick provisioning on the vMA.  I doubt this has any affect on the UPS communication but I could be wrong so I may try that approach today.  FYI, I had a hard time installing the VMWare tools but I finally got them working (somewhat).  VMware tools are for time synchronization and shutdown helping.  However, since the shutdown was working fine beforehand I don't think this is the issue.  This may be a time synchronization issue, so checked my time stamps on the vMA and the ESX.  They were way off so I fixed those and will try a new installation with all of those settings tonight.

I apologize for the long-winded explanation, but I need to have this UPS up and running by the end of the day on Thursday so I only have 2 more days to finish it!  Can anyone please share their wisdom and APC ninja skills with me?

Thank you very much.

[BillP · Accepted solution]

Hi,

It sounds like the HP server has iLO enabled. Check out these HP knowledge base documents.

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c00843255〈=en&cc=us&taskId=10...

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&objectID=c01139344&jumpi...

[Anonymous user]

Thank you for the reply, BIll.

I hadn't thought of that so I will try it when I get back to the office later today and report my findings.  You are correct that it has an iLO port and if that's blocking serial connections that may be our culprit.

Also to report my findings from last night's tests, I reset the APC unit again and matched the times up on the servers (manually) and left it overnight.  This morning when I checked it's still communicating!  I'll check once more when I get back to the office later today.

[Anonymous user]

Bill that was remarkably simple.  It seems to be working for now because right after I disabled CLI through the serial cable on the iLO the communication was back!  I will see if this is still the case tomorrow but that's amazing that it fixed things so quickly!  Thank you so much for the help

[voidstar_apc]

Definitely qualifies as ninja skills!