Who says it is running on top of Linux? I plan on using a Windows hypervisor.
I think Rob is on the right track. Signaling through serial or USB is not a viable way to do it anyway. For the size of a VM server, the UPS can be a "real" one that supports ethernet, and SNMP. Then you can have VM's all do their own thing.
Chuck Stickelman wrote:
Look, the Hypervisor is running ontop of Linux. From that perspective each Virtual Macine is a Linux process. What else could it be? If they are Linux processes then they will respond to Linux inter-process communications. What more is there for me to do? If you want, I can run a 'ps' command on authentic in ASG's lab and show you... The Windows/MAC OSs wn't see the SIGHUP. The Virtual Mahine layer will catch that and translate it into an OS-specific shutdown.This is exactly what Jim's scriCan you back up your statement that an instance of Windows NT appears to the hypervisor to be a Linux app? Even id the hypervisor is a Windows application? Doesn't make sense to me.I stand by my statement that a SIGHUP is meaningless to any OS other than Linux.Chuck Stickelman wrote:Writing a program that sends a SIGHUP to another process upon a specific event id trivial. I'm not programmer butr I could do it in a BASAH script. >From trhe hypervisor's point of view they are Linux apps and therefore should behaive as such. Linux apps should behaive properly when they catch a signal. The shell-script that was posted earlier effectively does just that. The good would be that the guest OS would initiate a shutdown upon seeing the SIGHUP from the Hypervisor... Sems pretty useful to me... Chuck -----Original Message----- From: "larry" <larry@xxxxxxxxxxxx> To: "ncolug@xxxxxxxxxxxxx" <ncolug@xxxxxxxxxxxxx> Sent: 11/23/2007 9:40 AM Subject: [ncolug] Re: vmI am quite sure I am missing things, the point of your message being only one of many... that is why I am running this by you all.1.) no one has verified that any hypervisor is ABLE to send a SIGHUP in response to an external event 2.) they are *not* "after all, Linux apps" - they are instances of operating systems 3.) if it did send it, what good would that do for a Windows guest? Or a MAC OS guest?Chuck Stickelman wrote:Larry, I think you mised the point of my prevoius message... The guests *should* shut down cleanly when the Hypervisor sends them a SIGHUP. (...regardless of why the Hypervisor sends the signal...) Sending a SIGHUP to the guest VM *should* be the same as choosing to shut it down from inside the VM... They are, afterall, just Linux apps... Of couse the UPS doesn't send the SIGHUP that's the job of the Linux Hypervisor. Which is the best choice for monitoring the UPS... Though some UPS monitoring applications have LAN awareness and can push the UPS status out to multiple systems - both Physical and Virtual... Chuck -----Original Message----- From: "larry" <larry@xxxxxxxxxxxx> To: "ncolug@xxxxxxxxxxxxx" <ncolug@xxxxxxxxxxxxx> Sent: 11/22/2007 10:32 PM Subject: [ncolug] Re: vmNo, the guests do NOT "shut down cleanly" just because you stopped or shut down the hypervisor.You want to talk about "standards" in regard to UPS signaling behaviour?! Which of the dozen or so standards would you like to discuss?Mike wrote:Chuck Stickelman wrote:So what I hear you say is: Hook the UPS to the physical machine Have the host OS monitor the UPS's state When the UPS signals the host it sends a SIGHUP to the Virtual Machines The VMs should then interpret the SIGHUP as a Shutdown command Is that what you had in mnd? Does anyone know if that's how it works? ChuckYes.The signal may depend on the VM design though. I would hope they have used enough sense to honor some standard.This really is a near trivial problem. The guests shutdown cleanly during a normal shutdown, yes? Why should a shutdown instituted by UPS software or even admin written (monitoring) scripts be any different?MikeTo unsubscribe send to ncolug-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field.
--"Perception is strong and sight weak. In strategy it is important to see distant things as if they were close and to take a distanced view of close things." Miyamoto Musashi (1584-1645)
To unsubscribe send to ncolug-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field.
