Committed on 4.0, 515. For 505, I'm not sure fix by Chip will work as logging 
into api.log happens by the servlet (APIServlet) some fix like this should work:

diff --git a/server/src/com/cloud/api/ApiServlet.java 
b/server/src/com/cloud/api/ApiServlet.java
index 8a1d4de..3ab6497 100755
--- a/server/src/com/cloud/api/ApiServlet.java
+++ b/server/src/com/cloud/api/ApiServlet.java
@@ -103,6 +103,13 @@ public class ApiServlet extends HttpServlet {
         }
     }
 
+    /*
+     * Strips off sensitive content based on
+     */
+    private String stripSensitiveContent(String str) {
+
+    }
+
     @SuppressWarnings("unchecked")
     private void processRequest(HttpServletRequest req, HttpServletResponse 
resp) {
         StringBuffer auditTrailSb = new StringBuffer();
@@ -334,7 +341,7 @@ public class ApiServlet extends HttpServlet {
                 auditTrailSb.append(" unknown exception writing api response");
             }
         } finally {
-            s_accessLogger.info(auditTrailSb.toString());
+            
s_accessLogger.info(stripSensitiveContent(auditTrailSb.toString()));
             // cleanup user context to prevent from being peeked in other 
request context
             UserContext.unregisterContext();
         }

Some work on refactoring the api layer is going on api_refactoring, the goal is 
to separate policy from mechanism and separate tightly coupled security checks 
using annotations, and also fix and automate docs. Because this the 
APIServlet.java will have a function, one point to strip out sensitive data 
like passwords and ssh-keys from logs instead of not logging them completely. 
I'll start another thread on api_refactoring and this issue.

Regards.

On 29-Nov-2012, at 9:34 AM, Chip Childers <chip.child...@sungard.com> wrote:

> On Wed, Nov 28, 2012 at 7:41 PM, Joe Brockmeier <j...@zonker.net> wrote:
>> On Thu, Nov 29, 2012 at 11:56:39AM -0500, Chip Childers wrote:
>>> I'll look at 505.
>> 
>> Great, thanks!
> 
> Fix committed to master and 4.0 branches (from the air no-less) ;-)
> 
> -chip

Reply via email to