Avoiding common pitfallsEdit online Common problems and how to avoid them. Ensuring correct channel path statusEnsure that you have varied the path offline before you perform a planned task on it.Determining channel path usage on LPARTo determine the usage of a specific channel path on LPAR, for example, to check whether traffic is distributed evenly over all channel paths, use the channel path measurement facility. Ignore unnecessary I/O devicesA Linux® LPAR should contain only those I/O devices that it uses. Using cio_ignoreWith cio_ignore, essential devices might be hidden.Excessive z/VM guest swappingAvoid excessive guest swapping by using the timed page pool size and the static page pool size attributes.Including service levels of the hardware and the hypervisorThe service levels of the different hardware cards, the LPAR level, and the z/VM® service level are valuable information for problem analysis.Booting stops with disabled wait stateAn automatic processor type check might stop the boot process with a disabled wait PSW.Auto-configuration overrides persistent configurationAfter upgrading Linux®, persistent device configurations that were specified with chzdev are overridden with the auto-configuration.Preparing for dump-on-panicSet up your system to automatically create a memory dump after a kernel panic.Multipath failover causes kernel panicIn a multipath setup where SCSI disks are attached over multiple paths, failover might trigger a kernel panic.Preventing excessive kernel parameters during installationThe Red Hat® Enterprise Linux® installation program automatically generates one rd.zfcp boot parameter for each FC-attached SCSI disk that is a dependency required to mount the root file system that it finds. This automation can lead to the kernel command-line parameters exceeding the limit of 4096 characters. Function unavailable or degraded in Linux on z/VMFor some functions, Linux® on z/VM® issues diagnose instructions to the z/VM® hypervisor.Parent topic: Diagnostics and troubleshooting
