tracker issue : CF-3726686

select a category, or use search below
(searches all categories and all time range)
Title:

Code changes not reliably seen if 'Cache Template in Request' is enabled

| View in Tracker

Status/Resolution/Reason: Closed/Withdrawn/

Reporter/Name(from Bugbase): Tim Parker / Tim Parker (Tim Parker)

Created: 03/19/2014

Components: Caching

Versions: 9.0

Failure Type:

Found In Build/Fixed In Build: 9.0 /

Priority/Frequency: Major / Most users will encounter

Locale/System: English / Platforms All

Vote Count: 0

Problem Description: code changes don't reliably take effect if the 'Cache Template in Request' setting is used.  This problem has been reported to me with both CF9 and CF10 (see test config) We were initially unable to reproduce this problem in a simple scenario, but then realized that the 'problem' modules were being called from within a CFThread block.  The two attached modules demonstrate the problem *sometimes*.  Our scenario is significantly more complex, and the problem occurs more frequently/reliably

Steps to Reproduce:
1) unzip the attached archive (two CF modules) in a CF-visible directory
2) browse to 'call-foo.cfm'
3) edit the assignment on line 1 of 'foo.cfm'
4) reload 'call-foo.cfm' in your browser - if this reload is soon after the updated 'foo.cfm' is saved, the old version of 'foo.cfm' is seen
  >>> NOTE: this is NOT a 100% reproducible case - the change is *sometimes* seen on the first reload, sometimes it takes a few reloads before the update is visible.


Actual Result: code change is not reliably reflected shortly after code is updated

Expected Result: code change should be reflected immediately - all requests started after the module update should see the update.  This is reliably the case with 'Cache Template in Request' disabled, but not with the setting enabled

Any Workarounds: turn off 'Cache Template in Request'

----------------------------- Additional Watson Details -----------------------------

Watson Bug ID:	3726686

External Customer Info:
External Company:  
External Customer Name: TParker
External Customer Email:  
External Test Config: My Hardware and Environment details:

CFML Engine Version: ColdFusion Server 9.0.2.282541 (JRun/4.0)

         JVM version: 1.7.0_51-b13 - [64 bits, Windows 7]

     JVM memory (MB):  max:996 total:996 free:756.41



-----  AND -------

 CFML Engine Version: ColdFusion Server 10.0.13.287689 (Apache Tomcat/7.0.23)

         JVM version: 1.7.0_51-b13 - [64 bits, Windows 7]

     JVM memory (MB):  max:954 total:954 free:123.96

Attachments:

Comments:

Hi, please attach the repro code for this. (Comment added from ex-user id:sandeepp)
Comment by Adobe D.
13031 | May 12, 2014 12:30:47 AM GMT
We don't have a 100% reproducible case for this, and have turned off 'cache template in request' on all development servers where this problem has occurred - debugging is difficult enough without being uncertain about whether your code changes are actually executing or not. CFThread is a possible contributor to the problem, although this is just a hunch - in general, we've found CFThread to be unstable (lost sessions, etc) with CF10, and as a result we have rewritten most of the uses which were working with CF9 so we don't have to debug CF itself. The only case where we're using CFThread in current code is to trap CFAbort in background jobs - in this case, we launch a thread and wait for it to complete before proceeding with anything else.
Comment by External U.
13032 | September 03, 2014 02:00:16 PM GMT
@TParker, in the bug description, you have asked to use the attached archive and browse a cfm. However, there is no file attached to the bug. Could you please attach it?
Comment by Rupesh K.
13033 | November 24, 2014 03:32:28 AM GMT
Please provide us the required files to repro the case.Thanks.
Comment by Akhila K.
13034 | December 03, 2014 12:44:11 AM GMT
We are closing this bug. Incase if user is still facing the issue and wants it to be fixed, please let us know. Also please provide us required information. We will look into it.Thanks.
Comment by Akhila K.
13035 | December 08, 2014 04:00:58 AM GMT