Kentico/Azure environmental issues.

Vic Carter asked on January 30, 2015 18:36

One of our clients is running Kentico 7 in an Azure environment. I do not have access to the actual environment and do all development via CMSDesk and SiteManager of which I have full admin privileges(so far as I can tell). Multiple sites are handled by this instance.

Here is a summary of several issues we have been having since development began two years ago.

When logged in to CMSDesk the Check Out/Check In process is inconsistent. Multiple users have reported instances where they could not update a webpart due to the system saying it was checked out by another user. Sometimes it is another user listed, sometimes it is the current user listed. Any attempts to refresh the page will show that the Checked Out/Checked In state will flip flop, going from one to the other seemingly randomly. This has occurred while viewing the Live Site as well, where changes would appear then disappear. Needless to say, this has caused debugging actual issues with development quite difficult. Though as annoying as this is, a more serious and troubling aspect of this issue(I feel they are related) shows up in the form of content that has been staged or has gone live, suddenly reverting to an earlier state, sometimes undoing weeks of work and forcing us to roll back or in some cases to rebuild large sections of the sites.

Every time something has been adjusted either at the request of MS or Kentico, other performance related problems have seemed to improve but this underlying issue remains.

ANY information people may have with similar issues or ideas as to what the cause may be, would be greatly appreciated.

Please ask for any clarification you may need.

Thank You

azure inconsistency kentico 7

Recent Answers

Brenden Kehren answered on January 30, 2015 19:27

I've encountered this in all my web farm instances no matter it it's on Azure or not. Seems to be a caching issue in most cases and I can get what I need working by clearing the cache on the server in Kentico. Downside is why do I have to do this? Shouldn't checking an object in automatically clear that cache so it shows current for other users?

I feel your pain Vic.

0 votesVote for this answer Mark as a Correct answer

Vic Carter answered on January 30, 2015 19:52

Thanks Brenden, it is actually nice to know I am not the only one.

0 votesVote for this answer Mark as a Correct answer

Filip Ligač answered on February 2, 2015 12:57

Hello Vic,

The symptoms you described indicate that web farm synchronization is not working properly on your Kentico instance. The thing is the server cache should be automatically flushed on all servers which are part of the web farm (all Azure instances in your case) everytime you make a change on any of the instances.

If that fails to happen, it is most probably caused by a bug and I am afraid there were several of them fixed in Kentico 7 related to web farm synchronization. With that being said, due to this reason I would strongly recommend that you apply the latest hotfix (7.0.103).

Let me know in a response if you still experience the same issues after updating.

Best regards,

Filip

0 votesVote for this answer Mark as a Correct answer

Vic Carter answered on February 2, 2015 15:46

Thank you Filip, I was on a call with Miro Remias on Friday debugging some problems and he noticed that this seemed to be a problem. I think our client is running the latest version of 7 with all of the hotfixes so they are looking at solutions right now. I appreciate your response and I am glad to have a more concrete understanding of what is likely to be causing the issue.

0 votesVote for this answer Mark as a Correct answer

Brenden Kehren answered on February 2, 2015 19:23

Interesting Vic, I have a version 7.0.40 instance running perfect without issue. Although the 8.1.2 instance I have in a similar environment doesn't run so well. I'd be interested to hear what you come up with for a solution.

0 votesVote for this answer Mark as a Correct answer