AMD Ryzen 7 3500 laptop (16 cores, 32 GB RAM) with VMs located on an external USB 3.0 2 TB SanDisk Extreme Pro v2 SSD. Everything was working fine, up until a Lenovo display/docking adapter USB driver update.
After that update, suspend performance was atrocious.
Windows 10 host system error logs showed multiple 51 (disk paging) and 153 (IO operation retries with code 00 04 2A) messages. CrystalDiskInfo found no drive issues whatsoever.
VMWare log:
2022-07-08T12:50:24.885Z In(05) vmx SUSPEND: Start suspend (flags=0)
2022-07-08T12:50:24.891Z In(05) svga MKSScreenShotMgr: Taking a screenshot
2022-07-08T12:50:25.262Z In(05) vcpu-0 Closing all the disks of the VM.
2022-07-08T12:50:25.262Z In(05) vcpu-0 Closing disk 'nvme0:0'
2022-07-08T12:50:25.273Z In(05) vcpu-0 Progress -1% (msg.checkpoint.saveStatus)
2022-07-08T12:50:25.273Z In(05) vcpu-0 Checkpointed in VMware Workstation, 16.2.3, build-19376536, Windows Host
2022-07-08T12:50:25.295Z In(05) vcpu-0 OBJLIB-FILEBE : FileBEIoctl: Failed to set sparse file, 'Incorrect function' (10004) on 'F:\Virtual Machines\RTLS2\RTLS2-d9b37b43.vmem' : Incorrect function (65538)
2022-07-08T12:50:25.295Z In(05) vcpu-0 MainMem: Failed to set sparse file, 'Incorrect function' (10004).
2022-07-08T12:50:25.295Z In(05) vcpu-0 MainMem: Write full memory image (SF=0 L=0 SZ=0 RM=0 U=0 TM=0).
2022-07-08T13:08:00.030Z In(05) vcpu-0 MainMem: Wrote 4096MiB in 1054735ms (4.07 MB/s)
Existing config.ini file was as follows, thanks to some prior tweaking:
.encoding = "UTF-8"
authd.client.port = "903"
authd.proxy.nfc = "vmware-hostd:ha-nfc"
installerDefaults.autoSoftwareUpdateEnabled = "yes"
installerDefaults.autoSoftwareUpdateEnabled.epoch = "6025"
installerDefaults.componentDownloadEnabled = "yes"
installerDefaults.dataCollectionEnabled = "yes"
installerDefaults.dataCollectionEnabled.epoch = "6025"
mainMem.useNamedFile = "false"
mainMem.writeZeroes = "true"
mainMem.ioBlockPages = "4096"
mainMem.iowait = "0"
prefvmx.useRecommendedLockedMemSize = "FALSE"
prefvmx.allVMMemoryLimit = "24035"
prefvmx.minVmMemPct = "100"
I finally found the apparently unrelated (?) KB below, regarding long shutdown times. Curiously, in this instance, there were no issues with VM shutdown, only suspend time. I appended the recommended config.ini settings, and suspend magically returned to normal performance.
https://kb.vmware.com/s/article/2030543
2022-07-08T16:53:53.999Z In(05) vcpu-0 MainMem: Wrote 4096MiB in 21941ms (195.75 MB/s)
