Does anyone have scripts that utilize something called the 'DD' command in order to backup/capture the first 25MB of a given VMFS datastore (aka the meta data?). I hear this is good practice, so support can (if/ever necessary) recover VMFS information in case of corruption? Thank you. (I'm admittedly, as you might tell, a total Linux neophyte).
You can use:
dd if=/vmfs/volumes/VMFSName of=NameOfFile bs=1048576 count=25
Or something like that. Where did you here that the VMFS metadata should be backed up? I am interested in how this would help during a forensic restoration.
Edward L. Haletky, author of the forthcoming 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', publishing January 2008, (c) 2008 Pearson Education. Available on Rough Cuts at http://safari.informit.com/9780132302074
I actually heard about this in a VMworld session entitled "Advanced VMFS Configuration and Troubleshooting" (TA48 is session ID). The last half of it had many people scratching their heads in bewilderment, and wondering if and when and how a given VMFS volume might become corrupt to the point where the volume's content just might be completely unusable. This created a few uneasy murmurs from the crowd. I hadn't heard about this happening prior to this revelation, but I just have to wonder: If this is possible, why doesn't VMware provide a script for their customers or otherwise provide built-in functionality to allow for this 'first 20-25MB' of VMFS metadata to be backed up, without customers having to write their own scripts, etc.?
Thanks for the feedback. Actually, when I looked up the title of this session, and its slides (presentation attached), I did in fact come across one of the slides showing how the dd command could be used. But I'm still not sure how far I'm going to pursue this.