Home
| Calendar
| Mail Lists
| List Archives
| Desktop SIG
| Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings Linux Cafe | Meeting Notes | Linux Links | Bling | About BLU |
Hello my fellow hackers! I am new to the list (and Boston) and am interested in OpenQRM as an open-source resource management solution possibility, since VMware ESX Server + VMware Lab Manager is mighty expensive and does not support Solaris (SPARC). Does anyone have experience with OpenQRM? I would possibly use Xen, but unfortunately, I don't believe that our server hardware (Intel Xeon) has virtualization extensions necessary to run Windows guests unmodified. My ultimate goal is to create a large physical infrastructure that can launch test targets "on-the-fly" and then instruct each guest to perform some type of test and report back the results, then when done, clean up the environment (revert to snapshot). I currently have ~64 physical servers each with 8GB of memory, 800 GB of storage (400GB x 2), Intel Xeons (3.2 GHz x 2), and two NICs (1000BT x 2). I also have set up some software-based iSCSI targets, but pointing this many hosts at one physical iSCSI disk causes massive slow down! Scenario: Simulate network of 100 Windows XP SP2, 20 Windows 2000 Pro SP4, 10 Red Hat AS 4.x, 5 Red Hat WS 3.x, and 2 Solaris 9 (SPARC) machines. A physically located central server, outside of this virtual infrastructure, will process requests and communicate with all of these virtual hosts. Execution: Log into some type of management interface (or shell script, api, etc) and instruct the virtual infrastructure to create a virtual configuration as described above. When all machines have booted, launch the test scripts and reports the results to some central server (or email them). Then revert the virtual infrastructure and/or undeploy the configuration from the physical hardware platforms. I am not sure if OpenQRM can even due this, as it appears to utilize PXE booting. This would mean that each physical server would only get one instance, but the server has the memory/cpu/hdd capability to run up to ~64 Operating Systems in parallel (assuming ~128MB / VM). Any suggestions are greatly appreciated ... thanks! -- Kristian Hermansen -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
BLU is a member of BostonUserGroups | |
We also thank MIT for the use of their facilities. |