6

We have been seeing weird things with one of our publisher nodes. We have two publisher nodes. One seems to be working well. When there are no publish jobs, nothing is written to the cd_transport log but on our other node, I keep seeing the following entry every second or so;

NODE1 (In question)

2013-06-21 10:48:23,677 TRACE RollingStateRetriever - Batch size:...0

2013-06-21 10:48:23,677 TRACE TransportService - Duration of notification batch fetching: 0

NODE2 (Expected entry)

2013-06-21 10:44:55,193 INFO HTTPSTransportConnector - Deleted file at remote HTTP(s) destination: tcm_0-315375-66560.state.xml was success: true

In addition, does anyone know where I can find a helpful document for the publishing process and troubleshooting?

Improve this question
1
7

If you have two apparently similar nodes that behave differently, the greatest suspicion always falls on permissions. For example, do both deployers/brokers run under similar accounts and are the various file system permissions set correctly for those accounts?

Improve this answer
2
  • They are using the same account – BryanD Jun 25 '13 at 21:36
  • That still leaves the permissions settings on each server. – Dominic Cronin♦ Jun 27 '13 at 22:00
5

There doesn't appear to be an error in the log entries that you have, so it may just be that the logging levels between the two nodes are inconsistent (and the same thing is happening on each).

The Publisher Service logging is enabled in the Tridion MMC snap-in (under Publisher settings). enter image description here

Once this is done, the location and logging level is done in the following location: [Tridion]\bin\TCMPublisher.exe.config

According to Elena's logging blog post, logging from the transport service is done in the [Tridion]\config\logback.xml file and output to the cd_transport.yyyy-mm-dd.log log file (by default).

<appender name="rollingTransportLog" class="ch.qos.logback.core.rolling.RollingFileAppender">
    <rollingPolicy class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
        <fileNamePattern>${log.folder}/cd_transport.%d{yyyy-MM-dd}.log</fileNamePattern>
        <maxHistory>${log.history}</maxHistory>
    </rollingPolicy>
    <encoder>
        <pattern>${log.pattern}</pattern>
    </encoder>
    <prudent>true</prudent>
</appender>

UPDATE [24/06/2013]:

It appears that the question has been renamed to "Two publishers: one deletes remote files, the other one doesn't".

Are you definitely sure that this is happening? If an item is published, and added to the publish queue, then only one of your publishing nodes will pick this up and process it. (Leaving the other node waiting for other published items to process.)

Your log file entries do not seem to suggest that there is an error removing the remote files. Are your files still present on the filesystem of the machine running the Deployer? If so, then, depending on what files are being left, that it is the Deployer failing to clean them up. You may want to check that the Cleanup attribute is set to true in the following node on your cd_deployer_conf.xml file:

<Queue>
    <Location Path="D:\Tridion\Incoming" WindowSize="20" Workers="20" Cleanup="true" Interval="30s"/>
</Queue>

Can you update the question with more details about exactly what the issues you are seeing are please?

Improve this answer
2
  • The only time cleanup does not happen is when I restart the Transport and publisher services due to jobs stuck in queue. I guess I am looking for a reference for the publishing process and troubleshooting. Not much in the SDL Tridion documentation on the entire process. THank you. – BryanD Jun 25 '13 at 21:35
  • As Nuno comments above, the Troubleshooting Content Distribution section of the online documentation (login required) is a good place to start. There is also a Troubleshooting scaled-out publishing section which may help with your multiple publisher issues too. – Jonathan Williams Jul 3 '13 at 9:04
1

I had faced a similar issue with our setup.

The setup was having two publishing nodes and both the nodes in the two different CDA servers were pointing to the same Tridion Broker Database.

I followed the steps mentioned in other answers above but it did not help. Then I came across this post by jowhnwinter: Multiple deployers single Content Delivery database (Broker DB) and the post by krishnakumar: SDL Tridion 2011 Unpublishing Issue in webfarm and decided to try out a few changes in the setup.

I created two instances of Tridion Broker Database and made changes to cd_storage_conf.xml like:

  1. Server1 was pointing to Tridion_Broker1
  2. Server2 was pointing to Tridion_Broker2

Surprisingly, the files started getting deleted from both the servers.

So, we settled for one publishing node on Server1 and a Symbolic Link Junction folder on Server2 pointing to the same database.

If its not too late, please make these changes and test the logs again.

Improve this answer

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.