Fix: Stop and clean finished/running openvas process before resumming a scan

If Ospd-Openvas crashes during a scan, probably the scan will continue
running in background, leaving the kb in redis occupied.
After ospd-openvas is initialized, the task can be resumed.
In this case, the scan id of the task will be the same one as before,
and there will be two kbs for the old and the new tasks with the same scanid.

This patch ensure that there is no kb in redis in used with the same scan id,
and also it will stop the scan in case it is still running.

This first commit add a function to check for kbs with the given scan id.

Author: jjnicola

ospd_openvas/db.py

@@ -648,6 +648,25 @@ def find_kb_database_by_scan_id(
 
         return None
 
+    def check_consistency(self, scan_id) -> Tuple[Optional[KbDB], int]:
+        """Check if the current scan id already exists in a kb.
+
+        Return a tuple with the kb or none, and an error code, being 0 if
+        the db is clean, -1 on old finished scan, -2 on still running scan.
+        """
+        err = 0
+
+        kb = self.find_kb_database_by_scan_id(scan_id)
+        current_status = None
+        if kb:
+            current_status = kb.get_status(scan_id)
+        if current_status == "finished":
+            err = -1
+        elif current_status == "stop_all" or current_status == "ready":
+            err = -2
+
+        return (kb, err)
+
     def release_database(self, database: BaseDB):
         self.release_database_by_index(database.index)
         database.flush()