2

Tridion Publisher is frequently down.This happens in the middle of publishing? Any ideas on how to debug the root cause? I restarted the Tridion Service Host in the content Manager server and the publisher comes up.

  • It is possible that your publisher can be down due to multiple reasons! what might help the cause is if you can provide some information to your question? Is your publisher scaled out i.e; do you have it running in the CM server or a different server? Does your publishing stall or fail as well? If your publishing is failing, what stage of publishing does it fail? You can look this up from the PublishingQueue? Do your publisher event or log show errors? Can you add it to the question? – Shiva Mar 10 '16 at 17:04
  • 2
    Another thing you can try is running the publisher in debug mode. Stop the windows service, and run tcmpublisher /debug - this will run it in a console and log messages, with a bit of luck you may get some info about the last operation that made it crash. Which version of Tridion are you running? – Nuno Linhares Mar 10 '16 at 17:05
  • Are there any customization like event system on publisher, if yes then disable those customization and try to monitor the publishing. We had similar issue when event system on publish was interfering with publishing and the service was getting stopped. – Hiren Kaku Mar 10 '16 at 17:24
  • When you say "Tridion publisher is frequently down" what do you mean? Other people have interpreted this as meaning the publisher service, but then it's unlikely that restarting the Service Host would fix this. Please describe the symptoms you are seeing. – Dominic Cronin Apr 28 '16 at 12:05
  • I think the general question of how to debug publishing is okay enough to leave open. I'd be looking for a general overview of how to troubleshoot publishing as an answer and I'm pretty sure we have this documented. – Alvin Reyes May 3 '16 at 8:11
2

This happens due to memory leakage, regarding memory leakage issues widely known for 5.3 and 2009 (mainly because of com+) as pointed by Pankaj.

We faced similar problem for one of the client and tried many a things by checking memory / garbage clear logic (disposable) in different areas. As a work around we placed a batch job (could be script or a .net program) scheduled which restarts the com+ and Publisher services after a certain time (we have monitored the tendency and time for the same).

This is not the desired solution but strange things happen around com+.

You can refer to this thread (though it was caused for event system and in your case it should not be responsible) just to find more information.

I suggest you need to find out the source for your memory leakage, maybe try putting some logs, check code (if Com / TDSE objects are released properly)

Please do mention the Tridion version for which you are facing issues and asking in the forum, it will help to provide the correct information and not based on any assumption.

Let me know if this helps !!

1

your question above is a bit abstract - it would help if you provide more information. anyway, I have faced this issue multiple times and following are few reasons/analysis we had done to resolve it:

  • The most probably cause of this happening (in our case) was a memory leak - this may be in the Deployer Extension, Storage Extension, Custom Resolver, TBBs etc. - In one instance of this error, we had this memory leak in the Deployer Extension and in another extension it was in the C# TBB in use
  • You have not mentioned the version of SDL Tridion - if it is legacy system like R5.3 or 2009 which heavily relies on the COM+ - then there is a huge possibilities of COM+ related errors - you must check and re-check your code for any possible optimizations
  • Another thing you should do is follow the SDL recommended Maintenance Plans for your SDL Installation - Reindexing, Purging etc. - and check it resolve your issue

Further, you should check in in the Windows Event Viewer on CM server or Deployer/Publisher logs on CD server for any possible Warning/Error logging. Also, you must try the Suggestion made by Nuno to stop the Publisher process and run it through command line in the debug mode as - tcmpublisher /debug. Monitor this and this would most probably gives you some meaningful error to track.

0

I had the same thing caused by an endless loop in the template. For us it was the main template, but only occurred when a metadata field linked to an item, and that item linked to the item with the metadata field... We added code to check if the 'linked to' item had already been added or not...

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.