Susan, > I'd like to save disk storage as well as simplify maintenance so thought > the concept of a basevol/guestvol penguin colony sounded perfect. Some shops have used it and are using it, I believe. Some have backed away from it. Are people still using basevol/guestvol?
> I'm a relative newbie, so what might be obvious to most administrators is > lost on me. Then perhaps you want to start with a simpler approach. >> I've had discussions about this with Mike before. We don't agree much. I recall one discussion, and yes we agreed to disagree. >> I believe the main reason Mike does it the way he does is that it simplifies the cloning process/scripts. Simplify, yes, but not really for the cloning script. Cloning 6 file systems instead of 1 should only require an additional 5 function calls to copy_disk(). The main reason we used just a single root file system is because there doesn't seem to be a "one size fits all" when trying to decide which file systems to break out and how much space to give them. This book was written to get shops going with z/VM and Linux. I know people who have used the books then modified "the master (golden) image" in a number of different ways. One just went to a single 3390-9 over / (to give 6.9GB of space instead of 2.3GB) and no modification to the clone script was needed. Another added more file systems per Mark's suggestion and modified the cloning process accordingly. The book quotes Albert Einstein: "Everything should be made as simple as possible, but not simpler." Perhaps a single file system is "too simpler" I know some shops are using a number o 3390-3s or maybe a single 3390-9 to create a single volume group. Perhaps 1/2 to 2/3 of that volume group is carved into logical volumes for file systems as Mark suggests. Then if any of the file systems start to fill up, the logical volumes can be dynamically extended. For a while on this list, there was a recommendation to keep the root file system on a conventional file system, not a logical volume. That can still be done while using logical volumes for all other non-root file systems. This is another, more flexible approach, that probably comes under the clause "but not simpler". I have been doing work with IBM middleware and find that /opt often needs to be large. I've modifed the cloning process to clone a two volume system: 100 minidisk over / and a 103 over /opt (101 and 102 are VDISK swap spaces). I wrote up that process in cookbook-style steps if anyone is interested. > The only way I've built a Linux system is outlined in the Redbook " z/VM > and Linux on zSeries: From LPAR to Virtual Servers in Two Days" There are updated books - see the top three entries on linuxvm.org/present - but they still use just a single root file system. What are other people using for the layout of their master/golden image? "Mike MacIsaac" <[EMAIL PROTECTED]> (845) 433-7061 ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
