Page tree
Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 14 Current »

Batches

If the batch cron job stops working

First, check the log file in log/whenever_cron.log to see if there are any error messages. If there aren't the lock file might be stuck. To ensure the batch job doesn't run more than one job at a time there is a lock file tmp/pids/batch_ingest.*. Delete the matching files and wait for the next run of the cron job.

Matterhorn

Ingest service reports as WARNING on Avalon's about/health page

When an ingest fails, it may leave the Matterhorn service in a bad state until subsequent ingests succeed. If you know that was not caused by a Matterhorn issue, you can bring its status back to OK by sanitizing the service:

1. Goto: http://MATTERHORN_URL:PORT
2. Login using credentials (IU: Test+server+documentation)
3. Goto: http://MATTERHORN_URL:PORT/docs.html?path=/services
4. Submit the sanitize form with serviceType=org.opencastproject.ingest&host=http://MATTERHORN_URL:PORT

Tomcat

Another Tomcat from a different repo was installed and replaced the old one

In Avalon 5.x and priors, we use a specific version of Tomcat from nul-repo. If a different version (for example one from epel) was installed, it will cause problems. To revert:

  • Move old logs: /usr/local/tomcat/logs /usr/local/tomcat/logs.old
  • Downgrade tomcat to 7.0.32-1.el6 from nul_public: sudo yum downgrade tomcat --disablerepo=epel

  • Restore server.xml, tomcat-users.xml and web.xml in fedora/WEB-INF with a working versions from github or another server

  • Restart Tomcat

Fedora

Corrupted thumbnail

Symptom

Item cannot be opened. Fetching the MasterFile in Rails console returns:

Ldp::HttpError: STATUS: 500 org.modeshape.persistence.relational.RelationalProviderException: java.nio.charset.MalformedInputException: Input length = 1

Remedy

The following steps attempt to reverse engineer and modify DB blobs. They should be tested thoroughly on a non-production replica first.

Stop Tomcat/Fedora.

Find the Modeshape uuid for the MasterFile using its noid (eg 4q77fs77w)

select * from MODESHAPE_REPOSITORY where content like '%4q77fs77w%';

Look in the result and find the last uuid, fcf96b47505d64ecc9ea96-31cf-49c4-b4e3-f4b193723378 in this case 

| fcf96b47505d644a95ff85-b6ab-4bdb-a150-6d5f21c9cda0 | 2017-06-08 01:17:35 | h  metadata @   id 3   fcf96b47505d644a95ff85-b6ab-4bdb-a150-6d5f21c9cda0  content   key 3   fcf96b47505d644a95ff85-b6ab-4bdb-a150-6d5f21c9cda0 parent 3   fcf96b47505d64454d94d5-2fb7-48d9-8847-90a8b0d79f07 properties �   http://www.jcp.org/jcr/1.0 �   createdBy 
                                                                 bypassAdmin primaryType    $name 
   nt:folder  created .   $date    2017-06-08T01:17:35.518-04:00  mixinTypes P   0 H   $name 8   {http://fedora.info/definitions/v4/repository#}Pairtree     children ]   0 U   key 3   fcf96b47505d64ecc9ea96-31cf-49c4-b4e3-f4b193723378 name 
   4q77fs77w   childrenInfo    count              |

Dump the MasterFile row using this uuid

mysqldump -u mysql_user -h mysql_host -p fedora_db --skip-add-drop-table --no-create-info --where="id = 'fcf96b47505d64ecc9ea96-31cf-49c4-b4e3-f4b193723378'" MODESHAPE_REPOSITORY > masterfile.sql

Edit masterfile.sql using vim and delete everything after poster through thumbnail

\0\0\0poster\0\0^C1\0U\0\0\0^Bkey\03\0\0\0fcf96b47505d643edfc8a7-a84d-46a8-ae6f-38767eed38d3\0^Bname\0\n\0\0\0thumbnail\0\0\0^CchildrenInfo\0^T\

Drop the existing DB row

delete from MODESHAPE_REPOSITORY where id='fcf96b47505d64ecc9ea96-31cf-49c4-b4e3-f4b193723378';

Insert the updated row back

mysql -u mysql_user -h mysql_host -p fedora_db < masterfile.sql

Start Tomcat/Fedora.

Open the item in a web browser and recreate the thumbnail using the player's built-in function.


  • No labels