Unknown Edition is Repeatedly Submitted Every 10 seconds and Terminates Abnormally

Our Rhythmyx Server Publishing Log is showing an unknown edition for an unknown site being run every 10 seconds, and then it’s status says it has “Terminated Abnormally”

See the attached image PublishLogScreenShot.jpg


The server.log file contains 1000’s of entries for the error

com.percussion.services.error.PSNotFoundException: Cannot find object id=‘1,485’ of type ‘Edition’

The Edition with id 1485 is an old Edition that has been deleted, but system keeps trying to run this edition every 10 seconds, this Edition is not scheduled to run under Scheduled Tasks.

When I view the record for the Unknown Edition in the Publishing Log, and click on the Job Id link to get more information on the failed job, Rhythmyx then reports the following error

Cannot get value for expression ‘#{sys_runtime_navigation.jobEditionData.name}’

javax.servlet.jsp.JspException: Cannot get value for expression ‘#{sys_runtime_navigation.jobEditionData.name}’ at org.apache.jsp.tag.web.layout.pubruntime_tag.doTag(org.apache.jsp.tag.web.layout.pubruntime_tag:112) at org.apache.jsp.ui.pubruntime.JobPubLog_jsp._jspx_meth_layout_pubruntime_0(org.apache.jsp.ui.pubruntime.JobPubLog_jsp:173) at

Has anyone ever experienced this problem before, if so, how did you fix it.

Thanks in advance

Tyrone

Did you ever find a solution for this? The same thing has just started happening on our site - a recently deleted edition tries and fails to run every 10 seconds.

Have you opened a TS incident for this issue?

FWIW, restarting the system made the problem go away.

Fascinating. I am curious to know if a record was in the publishing queue that could cause this “Unknown Edition.” I’m glad to know that restarting resolved your problem.

I am thinking this is either an old schedule that is being retried until it succeeds. Also you may have set up a publish on a workflow transition with a workflow effect. If this is the case and the transition is an aging transition, then if the item fails to transition it may retry every aging cycle which is 10 min by default. There may be some validation error on this 10min aging cycle in the server.log or the error may just be that this workflow effect is referring to an edition that does not exist.

We had/have this problem as well. It comes up every month or so for us. Yes, restarting does fix it, but we were told that the cause is that the publishes start running into each other. We have a few news related publishes that we push out every half hour. Now normally these finish well before the half hour is up, but if we run a full publish in the mean time, it would cause the edition to back up behind itself, and then for some reason these Unknown Editions start publishing every 10-50 seconds until the logs fill up. Then that would basically crash the server. I have a hard time clearing these logs out as well. The scheduled task for the log cleanup doesn’t seem to remove these.

Alex

Does the psx_qtriggers table contain any scheduled tasks that are not listed under Admin -> Scheduled Tasks ?

I’m back from holiday, and likewise this problem has come back. Sometimes a restart stops it for a while and sometimes it just stops on its own, only the start again a few days later.

There are references to an old, deleted, edition in PSX_PUBLICATION_STATUS and references to an old, deleted site in PSX_PUBLICATION_SITE_ITEM, and as one would expect the references match up to each other via PSX_PUBLICATION_DOC. So I can see why the editions are failing with “Unknown Edition” and “Unknown Site” messages, but I still can’t tell why the editions are running in the first place, let alone every 10 seconds. The PSX_QTRIGGERS table looks correct and all the real publishing jobs are finishing on time.

At the moment I can’t recreate the error on my dev system, so I’m going to copy my production system and see if I can get the error going on a system with no users. Once I’ve done that, and unless a solution leaps out at me, I’ll be sending a copy of the system to Tech Support.

I had an old “Publish Now” menu entry that pointed to an edition and site that no longer existed. Users must have been clicking it by accident, which kicked off the repeated publishing attempts.

Which is why it didn’t happen on my dev system - no users, hence no accidents. :wink:

So, fingers crossed, it’s fixed. But it would be nice if the system failed a bit more gracefully in cases like this.

Steve, great job on identifying the problem.

Seems like the publish.jsp could verify the existence of the specified site ID and edition ID before passing on the command to publish.

Update: Looks like it does try to verify the site ID in publish.jsp (v7.1) through

IPSGuid siteGuid = gmgr.makeGuid(Integer.parseInt(site), PSTypeEnum.SITE);

but the JSP doesn’t verify the edition ID; rather, it tries an automatic edition resolution, but if it fails it will fail back to the supplied edition id then proceed without verifying the ID.