Scheduled Publish

Vote:
 

Hey,

Recently my schedule publish has just stopped working.  Sometimes it invalidates on remote servers but the current server hasn't been invalidated and some of the post publish code hadn't completed running.  Looking at http://joelabrahamsson.com/how-episerver-cms-caches-pagedata-objects/ and other pages, they suggest that EpiServer delayed publish does not invalidate the cache on the other servers.  Is this correct? and if so does anyone know why?

#197816
Oct 15, 2018 8:36
Vote:
 

That post is very old. I'm pretty sure that all builtin scheduled jobs have been moved to ISychronizedObjectInstanceCache now, which works across servers. It seems the problem is something else ... 

#197818
Oct 15, 2018 9:30
Vote:
 

Well thats good and explains why I hadn't noticed it before.  Any ideas where I should look for the issue.  I don't seem to be able to find any related errors anywhere.

#197821
Oct 15, 2018 10:27
Vote:
 

Can't say for sure because it seems the problem is with your site configuration. If you edit a page (manually) on one server, do other servers get updated correctly? 

#197822
Oct 15, 2018 10:29
Vote:
 

Can't say for sure because it seems the problem is with your site configuration. If you edit a page (manually) on one server, do other servers get updated correctly? 

#197823
Oct 15, 2018 10:29
Vote:
 

Yep, everything works fine aside from delayed scheduled jobs.  Also sometimes the scheduled job says nothing published even though it has published some of it.  So I think something is crashing somewhere.

#197824
Oct 15, 2018 10:31
Vote:
 

@quan I am using Episer ver 9.8.2 , I tried to create my own  public class DelayedPublishJob : ScheduledJobBase to try and add some debug code but it uses internal contentrepository.  Is it possible to do this?

#197859
Oct 15, 2018 19:54
Vote:
 

Also we have the timeout set to a minute.  What would happen if it was still processing the last changes.  I can't see that the code handles that.

#197882
Oct 16, 2018 15:56
Vote:
 

So it seems it worked before and then stop working recently? You didnot mention if you are running on Azure or not, but you should go check the UDP setting (for on-premise) or Azure Events setting (for Azure). It sounds like if you run the job every minute it might send too many cache validation "events" across network that might have stuck somewhere 

#197930
Oct 17, 2018 13:27
Vote:
 

interesting would be also to see Remote Event page screenshot from Episerver DeveloperTools.

#197931
Oct 17, 2018 14:23
Vote:
 

Hey,

We are using Amazon SQS.  I checked and all the SQS queues were low.  Even though the job runs every minute, scheduled work is normally in batches and we are not using projects.  Is the remote events an Episerver Hosted thing?

#197942
Oct 17, 2018 16:57
Vote:
 

No it is not, it uses UDP/Azure Events/blah blah to broadcast messages 

#197943
Oct 17, 2018 16:58
Vote:
 

I appear to be confused or the fact we are on page two and the forum has displayed my initial message again has confused Quan.  Validis suggests there is page that displays the remote events.  Is that a thing?  I haven't found this, we use SQS and I have been monitoring the length of the queues from Amazon.

#197952
Oct 17, 2018 19:53
Vote:
 

I appear to be confused or the fact we are on page two and the forum has displayed my initial message again has confused Quan.  Validis suggests there is page that displays the remote events.  Is that a thing?  I haven't found this, we use SQS and I have been monitoring the length of the queues from Amazon.

#197953
Oct 17, 2018 19:53
Vote:
 

No I was answering your question, remote events are not a "Episerver hosted" thing. it uses the mechanism provided by the environment (WCF UDP, Azure Events or Amazon SQS - which I'm pretty clueless about) 

#197955
Oct 17, 2018 20:06
Vote:
 

ok, so yes we have remote events setup correctly and have been running for some time.  

In resposne to your statement:

 "It sounds like if you run the job every minute it might send too many cache validation "events" across network that might have stuck somewhere "

I don't believe this is the case as the SQS queue lengths in Amazon were empty and we don't normally have items scheduled for 4:01, 4:02, we normally schedule a lot of work for 4:00 and then there is not normally anything more for an hour.  We wanted the delay publish on 1 minute so we actually publish the work when the editor decides and not x amount of minues after.

In response to:

"So it seems it worked before and then stop working recently?"

Yes this is true, it randomly stopped working on Tuesday.  We increased the delay from 1 minute to 5 minutes and the problem seems to have vanished.

 

Valdis said:

"interesting would be also to see Remote Event page screenshot from Episerver DeveloperTools."

Does this mean there is a page somewhere in Episerver that shows me the remote events status on the server?

#197957
Oct 17, 2018 20:15
Vote:
 

was referrring to Episerver DevTools nuget pacakge. it has some interesting statistics around remote events.

#197959
Oct 17, 2018 21:44
Vote:
 

Hey,

Thanks for that.  I am afraid looking at https://nuget.episerver.com/package?id=EPiServer.DeveloperTools&v=2.1.0 I can not install as our CMS version is currently too low.

#197977
Oct 18, 2018 13:06
Vote:
 

Adam,
Note that several versions of that package are available. As you select a version, its list of Dependencies changes.

#198009
Edited, Oct 18, 2018 17:44
Vote:
 

the lowest available is for Episerver 7.5 - https://github.com/episerver/DeveloperTools/releases/tag/v1.1.0

#198019
Oct 19, 2018 4:46
Vote:
 

which epi version are you on btw?

#198074
Oct 19, 2018 17:02
Vote:
 

There are many packages available.No scarcity of options.As soon as u change package,ur dependencies change.Regards.

#201621
Feb 25, 2019 11:20
Vote:
 

I have exactly the same problem. Anyone got a suggestion?

#202100
Mar 14, 2019 9:57
Vote:
 

Hey,

I have a few bits of advice:

  1. Ensure the scheduler is only running on one machine. (https://world.episerver.com/forum/developer-forum/-Episerver-75-CMS/Thread-Container/2014/11/disable-schedule-jobs-on-a-specific-server-in-a-load-balanced-environment/)
  2. Ensure that machine is pushing notifications to the other machines to invalidate the cache. (https://world.episerver.com/documentation/Items/Developers-Guide/Episerver-CMS/75/Event-management/Event-management/)
  3. Ensure you don't have a rouge machine processing those jobs.

I hope that helps.

#202113
Mar 14, 2019 14:15
* You are NOT allowed to include any hyperlinks in the post because your account hasn't associated to your company. User profile should be updated.