Dell PowerEdge R750xa EMC PMem 200 Series Users Guide - Page 45
ESXi troubleshooting and maintenance
View all Dell PowerEdge R750xa manuals
Add to My Manuals
Save this manual to your list of manuals |
Page 45 highlights
ESXi troubleshooting and maintenance NOTE: Go through/var/log/vmkernel.log to see if there are any errors or warnings reported related to persistent memory. The log entries related to persistent memory would be appended with NVD and or IntelNVDimm. ● Configuring Intel PMem in memory mode shows a reduction of system memory in system BIOS. ○ For example, a system configuration with 4 * 128 GB Intel PMems, configured in memory mode provides a system memory of 506.8 GB instead of 512 GB. ○ It is for reserving space for metadata. ● PMem datastore is not mounted automatically in ESXi. ○ See /var/log/vmkernel.log and check for shut down counter related errors to DIMM handles. ○ When there is a critical error to any of the Intel PMems, datastore is not mounted by ESXi. ● ESXi boot takes longer when Intel PMems are configured as Non-interleaved App Direct. Dell EMC does not support Non-interleaved App Direct goals with ESXi. ● If PMems contain a critical error, the ESXi Host Client (HTML client) GUI does not respond when the user moves to Persistent Memory section under Storage. The host server is unresponsive when it reads the PMem region address range with the critical error. For more information see VMware KB 70661: https://kb.vmware.com/s/article/70661. ● In ESXi, Windows as a guest operating system does not display Storage Class Memory device in the device manager, for the PMem devices that are attached to the virtual machine (VM). Virtual BIOS for the VM hides the device from device manager. User can use the PowerShell cmdlets to view the PMem devices that are attached to the windows guest operating systems running on ESXi. ● esxtop utility in ESXi exposes some dummy performance counters when PMems are configured in the system. Counter names might be displayed as 'Cache hit', 'Cache miss', 'nBuffers', 'Flush interval'. These are dummy performance counters and no functionalities that are associated with it for now. ● ESXi displays more number of NUMA nodes on systems that are configured with PMems in App-direct mode. This is an expected behavior as the NUMA nodes are created for volatile and nonvolatile address ranges. The nonvolatile address ranges are displayed as 0 MB. ● ESXi logs a warning in the Vmkernel log Unable to register PMem file system for APD notifications. This log entry can be safely ignored as the PSA (Pluggable Storage Architecture) APD(All Paths Down) plug-in is not supported for PMem type volumes. For more information see VMware KB 2145444: https://kb.vmware.com/s/article/2145444. ● When persistent memory goals are changed in sequence from App Direct mode to Memory mode and back to App Direct mode, PMem datastore does not mount automatically in ESXi. This occurs when ESXi is unable to format and mount the Namespaces when the App Direct mode is re-created. As a solution, perform sanitization operation before creating a new goal. VMware ESXi 45