>>> On 7/7/2014 at 10:02 PM, Cameron Seay <[email protected]> wrote: > Mark: > > That is what I call very cool. Let me be clear, DIRM finds available DASD > from the pool (in our case MOD 9s) and gives them to the clone? For > example, the clones will have volumes E150, E152... E15n, which are all > named the same for each clone anyway, and find DASD for them. I want to > make sure I'm clear on this.
Yes. There are some caveats, of course. A minidisk must be contiguous cylinders. So, if the free space in the pool gets fragmented, DIRMAINT might not be able to find enough contiguous cylinders to give to the new guest. In the Bad Old Days (before Linux 2.4), this was more of a problem because 1. Linux for System z wasn't able to have more than 1 partition per disk. 2. Logical Volume Manager wasn't available yet to make it easier to parcel out space within the guest. To avoid wasting disk space, people would allocate various sized minidisks to a guest to hold the various file systems, which definitely increased the possibility of fragmentation. These days, most shops just give guests entire full pack minidisks (1-END) and use partitioning and LVM within the guest to distribute the space. If you find that you're giving guests multiple 3390-9 volumes (for whatever reason), you might want to consider using SCSI over FCP for the non-OS data on those guests. You should be able to get a reasonable system installed on just one 3390-9, counting just the OS. Mark Post ---------------------------------------------------------------------- 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 ---------------------------------------------------------------------- For more information on Linux on System z, visit http://wiki.linuxvm.org/
