Internal Server Error in KMC after 11.0.4 Upgrade

Community Forums Backend Server Support
1

Hello,

After upgrading to CE 11.0.4 tonight, things seem to be going smoothly except for:

  1. DB Update script throws an error with no information. sql_updates file is present under /opt/kaltura/app/deployments , seems to contain SQL statements dated from 2014, nothing more recent. Not sure if this is OK (Upgrading from 10.21 CE)

  2. an error being returned when clicking on any content entries in the KMC. As the metadata loads for the entry, we are getting “Internal Server Error Occurred”. If you click “OK” it loads the rest of the entry information (description, etc.) and can play the video. Seeing this in the Kaltlogs on one of the front end nodes:

    2015-12-14 20:52:11 [0.000964] [10.2.208.5] [2125724345] [112] [API] [KalturaStatement->execute] DEBUG: Sql took - 0.00071811676025391 seconds
    2015-12-14 20:52:11 [0.000443] [10.2.208.5] [2125724345] [113] [API] [kCoreException->__construct] ERR: exception ‘kCoreException’ with message ‘Dynamic enum not found [transcript.Transcript] for type [assetType]’ in /opt/kaltura/app/alpha/lib/kPluginableEnumsManager.php:167
    Stack trace:
    #0 /opt/kaltura/app/plugins/content/transcript/TranscriptPlugin.php(84): kPluginableEnumsManager::apiToCore(‘assetType’, ‘transcript.Tran…’)

    #10 {main}
    2015-12-14 20:52:11 [0.000428] [10.2.208.5] [2125724345] [114] [API] [KalturaFrontController->getExceptionObject] CRIT: exception ‘kCoreException’ with message ‘Dynamic enum not found [transcript.Transcript] for type [assetType]’ in /opt/kaltura/app/alpha/lib/kPluginableEnumsManager.php:167
    Stack trace:
    #0 /opt/kaltura/app/plugins/content/transcript/TranscriptPlugin.php(84): kPluginableEnumsManager::apiToCore(‘assetType’, ‘transcript.Tran…’

These errors coincide when the error is received in the KMC. Any help appreciated

EDIT

Just wanted to add, kaltura-sanity.sh is passing on both front nodes, batch nodes, and the sphinx node.

2

Hi @siactive,

The issue starts here:

Did you run the config scripts post updating the RPMs?
I would suggest you run them again, after making sure $BASE_DIR/app/base-config-generator.lock and $BASE_DIR/app/base-config.lock do not exist.

The error above should be resolved by running:
php $BASE_DIR/app/generator/generate.php
and:
$APP_DIR/deployment/base/scripts/installPlugins.php

php $BASE_DIR/app/generator/generate.php should run as part of the postinst of kaltura-base, if it failed, try running it manually and check the results
php $BASE_DIR/app/generator/generate.php runs as part of the kaltura-base-config.sh script.

Before running these, you should stop Apache and Memcache, which kaltura-base-config.sh takes care of. If you do it manually, stop these daemons and also the monit watchdog so that it does not try to restart them.

3

Hi Jess,

I did run the post-install scripts on the nodes, so this morning I reran the base-config and others after removing the .lock files. While it completed successfully, I was still getting the error in the KMC.

So I shutdown the httpd, memcached and kaltura-monit services, and ran the PHP scripts manually. Turned everything back on, and the error has gone away, and I’m not seeing any additional errors in the kaltlog. Thank you for your help!

4

Sure, most welcome:)
My guess is something went wrong during the installPlugins.php run triggered by the kaltura-base’s RPM postinst. For instance, if the DB was inaccessible at that moment or something like that…

Anyhow, glad we’re good now.