update log colletion instructions to include 'previous' contain/pod instance logs

(cherry picked from commit 795dc78)

Author: Ariel Kass

docs/content/troubleshooting/log_status_collect.md

@@ -60,35 +60,39 @@ If the PVCs are not in the _Bound_ state, collect the events of all unbound PVCs
 
 #### Log collection for all CSI driver node pods and their containers
 
-To collect logs for all CSI driver node pods, use the following commands:
+To collect logs for all CSI driver node pods, use the following commands (collection of 'previous' pod logs may fail if the pods have never restarted):
 
     nodepods=`kubectl get pods -n <namespace> -l product=ibm-block-csi-driver -l app.kubernetes.io/component=csi-node --output=jsonpath={.items..metadata.name}`
 
     for pod in $nodepods;do for container in `kubectl get -n <namespace> pod $pod -o jsonpath='{.spec.containers[*].name}'`;do kubectl logs -n <namespace> $pod -c $container > logs/${pod}_${container}.log;done;done
+    for pod in $nodepods;do for container in `kubectl get -n <namespace> pod $pod -o jsonpath='{.spec.containers[*].name}'`;do kubectl logs --previous -n <namespace> $pod -c $container > logs/${pod}_${container}.previous.log;done;done
 
 <br>
 
 #### Log collection for all CSI controller containers
 
-To collect logs for all controller containers, use the following commands:
+To collect logs for all controller containers, use the following commands (collection of 'previous' pod logs may fail if the pod has never restarted):
 
     for container in `kubectl get -n <namespace> pod ibm-block-csi-controller-0 -o jsonpath='{.spec.containers[*].name}'`;do kubectl logs -n <namespace> ibm-block-csi-controller-0 -c $container > logs/ibm-block-csi-controller-0_${container}.log;done
+    for container in `kubectl get -n <namespace> pod ibm-block-csi-controller-0 -o jsonpath='{.spec.containers[*].name}'`;do kubectl logs --previous -n <namespace> ibm-block-csi-controller-0 -c $container > logs/ibm-block-csi-controller-0_${container}.previous.log;done
 
 <br>
 
 #### Log collection for CSI operator logs
-To collect CSI operator logs, use the following commands:
+To collect CSI operator logs, use the following commands (collection of 'previous' pod logs may fail if the pod has never restarted):
 
     operatorpod=`kubectl get pods -n <namespace> -l app.kubernetes.io/instance=ibm-block-csi-operator --output=jsonpath='{.items..metadata.name}'`
     kubectl logs $operatorpod -n <namespace> > logs/operator.log
+    kubectl logs --previous $operatorpod -n <namespace> > logs/operator.previous.log
 
 <br>
 
 #### Log collection when using host definer
-When a host definer issue occurs, be sure to collect logs from the host definer pod where the error occurred.
+When a host definer issue occurs, be sure to collect logs from the host definer pod where the error occurred, using the following command (collection of 'previous' pod logs may fail if the pods has never restarted):
 
     hostdefinerpod=`kubectl get pods -n <namespace> -l app.kubernetes.io/component=hostdefiner --output=jsonpath='{.items..metadata.name}'`
     kubectl logs $hostdefinerpod -n <namespace> > logs/hostdefiner.log
+    kubectl logs --previous $hostdefinerpod -n <namespace> > logs/hostdefiner.previous.log
 
 <br>