On 1/7/21 1:58 PM, Cohen, Sam wrote: > If you're running under z/VM with a class G (or lower) user, why use > cio_ignore at all? Your hypervisor will only allow you to see what should be > seen. Even without z/VM, you can limit the devices visible to the LPAR via > IOCDS statements. I don't have cio_ignore on the startup of any of my RHEL > guests (although I do use dasd.conf and zfcp.conf).
I, and others, have been saying the same thing for years. This feature was initially implemented simply because when people were doing an installation in an LPAR, booting was taking too long. It was taking too long because many customers make all I/O devices available to all their LPARs, and those customers tend to have a *lot* of I/O devices. The kernel can take quite some time to probe them all. When SUSE first shipped that feature, we made the mistake of inserting cio_ignore into the kernel parms by default. That caused problems like this one, and we changed the installer so that installs performed in a virtualized environment did not insert cio_ignore. I.e., z/VM and KVM. That eliminated a large number of customer problems. Mark Post ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www2.marist.edu/htbin/wlvindex?LINUX-390
