Improve the serialized SOAP error in CnsNodeVmAttachment/CnsNodeVmBatchAttachment #3663
Description
Context
VMOP consumes CnsNodeVmAttachment and CnsNodeVmBatchAttachment, sometimes the error returned inside contains serialized SOAP error, which contains things like OpIds and pointers that change on every failed reconcile attempt.
The OPID and pointers are not that useful, and if we simply add the content to VM.status.volume, then it cause unnecessary object update in every reconcile. Currently VMOP has following workaround to sanitize the error message. https://github.com/vmware-tanzu/vm-operator/blob/main/controllers/virtualmachine/volume/volume_controller.go#L1006-L1017
Example error message:
failed to attach cns volume: \"88854b48-2b1c-43f8-8889-de4b5ca2cab5\" to node vm: \"VirtualMachine:vm-42
[VirtualCenterHost: vc.vmware.com, UUID: 42080725-d6b0-c045-b24e-29c4dadca6f2, Datacenter: Datacenter
[Datacenter: Datacenter:datacenter, VirtualCenterHost: vc.vmware.com]]\".
fault: \"(*vimtypes.LocalizedMethodFault)(0xc003d9b9a0)({\\n DynamicData: (vimtypes.DynamicData)
{\\n },\\n Fault: (*vimtypes.ResourceInUse)(0xc002e69080)({\\n VimFault: (vimtypes.VimFault)
{\\n MethodFault: (vimtypes.MethodFault) {\\n FaultCause: (*vimtypes.LocalizedMethodFault)(\u003cnil\u003e),\\n
FaultMessage: ([]vimtypes.LocalizableMessage) \u003cnil\u003e\\n }\\n },\\n Type: (string) \\\"\\\",\\n Name:
(string) (len=6) \\\"volume\\\"\\n }),\\n LocalizedMessage: (string) (len=32)
\\\"The resource 'volume' is in use.\\\"\\n})\\n\". opId: \"67d69c68\""
Expectation
It would be best if CSI could sanitize the error message and remove the unnecessary opId/pointer. It might be arguable whether these things are necessary though.