Thanks for the info, I was wondering where the actual shutdown script was run from. When I get an outage window I will test again and see what the logs say.
I've attached what the /group1/bin/shutdown script currently looks like, do I need to worry about entering the username/password, or is that intentionally blank for scripting purposes?
PCNS 3.0.1 uses VIFPTARGET/vMATarget functionality provided by the vMA to store the ESXi host credentials. PCNS retrieves the credentials when connecting to the host to shut it down using the RCLI/Perl SDK. These credentials are actually stored in an XML file on the vMA. When you add a target host using the vifp addserver command a new accounts is created. We use the same file / grab the same info when running shutdownvms.sh
Again, thanks for the clarification Bill.
So I tested just running the shutdown script in /opt/APC/PowerChute/group1/bin/ and was getting the following error
So I made a backup of shutdown, and changed the line inside the script from:
perl ./bin/shutdownhost.pl --operation shutdown --username xxxxx --password xxxxx --loglevel debug
perl ./shutdownhost.pl --operation shutdown --username xxxxx --password xxxxx --loglevel debug
You received that error because you were attempting to run the shutdown file from within /opt/APC/PowerChute/group1/bin If you want to test the file without editing it you should cd to /opt/APC/PowerChute/group1 and then type sudo bin/shutdown
By editing the file as you have PowerChute will not shut the host down properly. You need to put the file back to its original content!
Yea in testing the script worked, but actually initiating a PowerChute shutdown still didn't shut down the hosts. I've recovered the original shutdown file so it should be default again. Where are the verbose logs stored for the shutdown file in /opt/APC/PowerChute/group1/bin/ ? I've specified the --loglevel debug in shutdown and I'd like to see exactly why these hosts still aren't restarting.
PCNS does not restart the host. If you want the host to restart you would need to configure that in the BIOS.
As for the debug log it is put in /tmp
I'm having the same issue I've seen floating around for a few other people, I'm able to get VM shutdowns to work in any order I want through a shell script, but the ESXi hosts I have this installed on won't work.
I've tested this on both ESXi 4.1 & ESXi 5.1 servers and I'm experiencing the same issue with PowerChute 3.0.1. Some are HA, and some are just regular standalone hosts, so it seems to be something other than the version of ESXi host causing the issues. One thing that does tie across all hosts is that they have symbols in their passwords, not sure if that would make a difference.
I've verified that all hosts and vCenter are licensed.
The host is added using vifptarget -s (vMA host always last in the list)
The host returns all NIC info from vicfg-nics --list command
I've added the export PERL_LWP_SSL_VERIFY_HOSTNAME=0 command to my shutdown script per FA159783
I don't believe this is the DNS issue in FA166844, since vifp listservers shows me the ESX hosts IP, not DNS name.
Again, VMs shut down without issue in whatever order I prefer, just the hosts I'm having issues with.
You can test PowerChute shutdown of the host manually. CD to /opt/APC/PowerChute/group1 (this is assuming app is installed to the default path) The command is sudo bin/shutdown If that does not work you can run the shutdown in debug mode by editing /opt/APC/PowerChute/group1/bin/shutdown (the instruction for editing the script are included in the file).
Choose a location