Implement the run collect-info through Edgeview

- Add the implementation on running collect-info through Edgeview
- Egeview is mostly only with read permission access to device filesystems
  and it does not want to import all the pkgs and replicate the collect-info.sh
  operations and maintain compatibility of two sets of the collect-info
- this implementation has a protocol to setup a request for debug container
  to start run 'collect-info.sh', and remove the request file after it is done;
  Edgeview then uses existing copy file protocol to transfer the collect-info
  tarball file back to user's laptop; then Edgeview put another request for
- debug container to remove the created tarball file in /persist
  debug container has added a background task to monitor the request-
- Since collect-info is mostly live info, there is no need to get all the newlogs,
  only fetch the newlogs of the past 10 days.
- change the debug/ssh.sh into debug/debug-tasks.sh to reflect it does
  not only handling the ssh task

Signed-off-by: Naiming Shen <naiming@zededa.com>

pkg/debug/Dockerfile

@@ -110,7 +110,7 @@ RUN mkdir -p /out/var/empty
 
 # tweaking various bit
 WORKDIR /out
-COPY ssh.sh spec.sh scripts/ ./usr/bin/
+COPY debug-tasks.sh spec.sh scripts/ ./usr/bin/
 RUN mkdir -p ./etc/ssh ./root/.ssh
 RUN chmod 0700 ./root/.ssh
 RUN cp /etc/passwd /etc/group ./etc/
@@ -126,4 +126,4 @@ WORKDIR /
 
 COPY --from=build /out/ /
 
-CMD ["/sbin/tini", "/usr/bin/ssh.sh"]
+CMD ["/sbin/tini", "/usr/bin/debug-tasks.sh"]

pkg/debug/debug-tasks.sh

@@ -0,0 +1,62 @@
+#!/bin/sh
+
+# setting things up for being able to access linux kernel symbols
+echo 0 >  /proc/sys/kernel/kptr_restrict
+echo -1 > /proc/sys/kernel/perf_event_paranoid
+
+# This script used to be 'ssh.sh', at the time of introducing the
+# monitoring task, it was renamed to 'debug-tasks.sh' to reflect
+# multiple tasks that debug container can perform in the background.
+
+KEYS=$(find /etc/ssh -name 'ssh_host_*_key')
+[ -z "$KEYS" ] && ssh-keygen -A >/dev/null 2>/dev/null
+
+# Background monitoring non-ssh related tasks
+monitor_file_and_execute_tasks() {
+    # edgeview request to generate collect-info and remove tar.gz file
+    # the protocol is for edgeview to create a file
+    # /run/edgeview/edgeview-request-collect-info, and debug container
+    # run 'collect-info.sh' to generate the tar.gz file, when the job is
+    # done, remove the file /run/edgeview/edgeview-request-collect-info.
+    # edgeview can also request to remove the tar.gz file by creating
+    # /run/edgeview/edgeview-request-remove-tar.gz file, and debug container
+    # will get the file name and remove the generated tar.gz file.
+    #
+    # if addition tasks are needed, this can monitor multiple directories
+    # and perform different actions by debug container in the future.
+    while true; do
+        if [ -f "/run/edgeview/edgeview-request-collect-info" ]; then
+            echo "edgeview request to run collect-info..." > /dev/kmsg
+            # the newlog part of the collection, only collect last 10 days
+            /usr/bin/collect-info.sh -t 10
+            # remove the request file
+            echo "edgeview request collect-info done" > /dev/kmsg
+            rm /run/edgeview/edgeview-request-collect-info
+        fi
+        if [ -f "/run/edgeview/edgeview-request-remove-tar-gz" ]; then
+            fileToRemove=$(cat /run/edgeview/edgeview-request-remove-tar-gz)
+            echo "edgeview request to remove $fileToRemove file" > /dev/kmsg
+            if [ -f "$fileToRemove" ]; then
+                rm "$fileToRemove"
+            fi
+            # remove the request file
+            rm /run/edgeview/edgeview-request-remove-tar-gz
+        fi
+        sleep 5
+    done
+}
+
+# Start the monitoring task in the background
+monitor_file_and_execute_tasks &
+
+if [ -f "/config/remote_access_disabled" ]; then
+    # this is picked up by newlogd
+    echo "Remote access disabled, ssh server not started" > /dev/kmsg
+    while true; do
+        # sleep for INT_MAX, keep the container running
+        sleep inf
+    done
+else
+    exec /usr/sbin/sshd -D -e
+fi
+

pkg/debug/scripts/collect-info.sh

@@ -6,7 +6,7 @@
 
 # Script version, don't forget to bump up once something is changed
 
-VERSION=24
+VERSION=25
 # Add required packages here, it will be passed to "apk add".
 # Once something added here don't forget to add the same package
 # to the Dockerfile ('ENV PKGS' line) of the debug container,
@@ -20,6 +20,7 @@ INFO_DIR="eve-info-v$VERSION-$DATE"
 TARBALL_FILE="/persist/$INFO_DIR.tar.gz"
 SCRIPT_DIR=$(dirname "$(readlink -f "$0")")
 
+READ_LOGS_DAYS=
 READ_LOGS_DEV=
 READ_LOGS_APP=
 TAR_WHOLE_SYS=
@@ -47,11 +48,12 @@ usage()
     echo "Read-logs mode:"
     echo "       -d                   - read device logs only"
     echo "       -a APPLICATION-UUID  - read specified application logs only"
+    echo "       -t NUMBER-OF-DAYS    - read logs from the last NUMBER-OF-DAYS [1-30]"
     echo "       -j                   - output logs in json"
     exit 1
 }
 
-while getopts "vhsa:dj" o; do
+while getopts "vhsa:djt:" o; do
     case "$o" in
         h)
             usage
@@ -63,6 +65,13 @@ while getopts "vhsa:dj" o; do
         a)
             READ_LOGS_APP="$OPTARG"
             ;;
+        t)
+            READ_LOGS_DAYS="$OPTARG"
+            if [ "$READ_LOGS_DAYS" -lt 1 ] || [ "$READ_LOGS_DAYS" -gt 30 ]; then
+                echo "Error: READ_LOGS_DAYS must be between 1 and 30."
+                exit 1
+            fi
+            ;;
         d)
             READ_LOGS_DEV=1
             ;;
@@ -157,6 +166,7 @@ fi
 # Create temporary dir
 echo "- basic setup"
 TMP_DIR=$(mktemp -d)
+LOG_TMP_DIR="$TMP_DIR/dayslogs"
 DIR="$TMP_DIR/$INFO_DIR"
 mkdir -p "$DIR"
 mkdir -p "$DIR/network"
@@ -433,9 +443,17 @@ find /sys/kernel/security -name "tpm*" | while read -r TPM; do
     fi
 done
 
+if [ -n "$READ_LOGS_DAYS" ]; then
+    mkdir -p "$LOG_TMP_DIR"
+    # Find and copy log files from /persist/newlog to $LOG_TMP_DIR in previous days
+    find /persist/newlog -type f -mtime -"$READ_LOGS_DAYS" -exec ln -s {} "$LOG_TMP_DIR" \;
+    ln -s "$LOG_TMP_DIR" "$DIR/persist-newlog"
+else
+    ln -s /persist/newlog "$DIR/persist-newlog"
+fi
+
 ln -s /persist/status       "$DIR/persist-status"
 ln -s /persist/log          "$DIR/persist-log"
-ln -s /persist/newlog       "$DIR/persist-newlog"
 ln -s /persist/netdump      "$DIR/persist-netdump"
 ln -s /persist/kcrashes     "$DIR/persist-kcrashes"
 [ -d /persist/memory-monitor/output ] && ln -s /persist/memory-monitor/output "$DIR/persist-memory-monitor-output"

pkg/edgeview/src/basics.go

@@ -94,6 +94,7 @@ func initOpts() {
 		"watcher",
 		"zedagent",
 		"zedclient",
+		"zedkube",
 		"zedmanager",
 		"zedrouter",
 		"zfsmanager"}
@@ -115,6 +116,7 @@ func initOpts() {
 		"app",
 		"configitem",
 		"cat",
+		"collectinfo",
 		"cp",
 		"datastore",
 		"dmesg",
@@ -686,6 +688,8 @@ func printHelp(opt string) {
 		// system
 		case "configitem":
 			helpOn("configitem", "display the device configitem settings, highlight the non-default values")
+		case "collectinfo":
+			helpOn("collectinfo", "collect the device information using collect-info.sh and download a compressed file in tar.gz format")
 		case "cp":
 			helpOn("cp/<path>", "copy file from the device to locally mounted directory by specify the path")
 			helpExample("cp//config/device.cert.pem", "copy the /config/device.cert.pem file to local directory", true)

pkg/edgeview/src/copyfile.go

@@ -37,6 +37,7 @@ const (
 	copyTarFiles
 	copyTechSupport
 	copyKubeConfig
+	copyCollectInfo
 )
 
 var (

pkg/edgeview/src/edge-view.go

@@ -230,6 +230,9 @@ func main() {
 		} else if strings.HasPrefix(pqueryopt, "tar/") {
 			psysopt = pqueryopt
 			fstatus.cType = copyTarFiles
+		} else if strings.HasPrefix(pqueryopt, "collectinfo") {
+			psysopt = pqueryopt
+			fstatus.cType = copyCollectInfo
 		} else {
 			_, err := checkOpts(pqueryopt, netopts)
 			if err != nil {