On 4/12/12 2:36 PM, Christopher Brooks wrote:
2. The JVM crashes. This does leave the standard java crash log.
There is no information in the matterhorn logs.
This usually happens when gstreamer dies. Can you post a
configuration file somewhere?
Which config file would you like?
The matterhorn config file, so I can see what you are trying to encode
at. Greg/Adam, do you know how to get the gstreamer error? (this
would require not running as a service)
Config file is attached.
This actually happened this morning. It started at about 4am or so
and continued until I restarted matterhorn around 10am. I monitor
load average and cpu utilization via snmp at 5 minute intervals. The
average utilization for this entire period was 1%. The graph is in
fact mostly flat. The load average for this period was less than 0.1.
If you are doing lots of little back to back recordings this could be
zip making disk contention an issue. What regiment of tests are
you throwing at it?
I am not doing lots of little back to back recordings. The recordings
are typically 1 hr to 1.5 hrs in length. There is a 5 minute break
between each recording. Basically the tests consist of scheduling 1 or
1.5 hr recordings of video, presentation, and audio captures throughout
the day. No recording was scheduled at 4am.
I know it isn't an awesome load average, but it's odd and it grows.
It's also considerably higher than normal for the agent when it is
not recording. Typically the load average is almost 0. It is the
vga2usb and the bluecherry card. We are using the -exact- reference
hardware that has been on the wiki for the last year.
Interesting. When not recording the load should be negligible. Maybe
something changed in confidence monitoring. Akm, thoughts?
Confidence monitoring is disabled as per the default configuration.
Furthermore I have already created a ticket about confidence monitoring
stating that enabling it crashes the JVM.
Chris
###########################################
### Matterhorn configuration properties ###
###########################################
# The HTTP server port. If you set this to port 80, you need to run Matterhorn
as root. Alternatively, if you want
# users to access Matterhorn on port 80 but do not want to run as root, keep
the default port (8080) and use an Apache
# HTTPD server with mod_proxy to forward port 80 traffic to Matterhorn on port
8080.
org.osgi.service.http.port=8080
# Whether Matterhorn itself should handle HTTPS traffic. Even if you set this
to 'false',you can still use an Apache
# HTTPD server as a proxy to handle SSL)
org.osgi.service.http.secure.enabled=false
# The secure server port to use if running Matterhorn itself with HTTPS (as
opposed to a proxy handling HTTPS).
#org.osgi.service.http.port.secure=8443
# The public URL of this matterhorn installation. If this felix installation
is proxied behind an Apache HTTPD
# reverse proxy, this URL should point to the proxy's port (usually 80).
org.opencastproject.server.url=http://xxx.xxx.xxx.xxx:8080
# The url of the remote service registry. This is used in cases where there is
no direct connection to the service
# registry database such as capture agens running in protected environments.
This is typically true for capture agents
# and should be set to the url of a server running the actual implementation of
the service registry and the path to
# the service registry(admin, worker, etc. See the build profiles in pom.xml
for a complete list).
org.opencastproject.serviceregistry.url=http://xxxx.media.berkeley.edu/services
# The base URL of the server hosting the administrative tools. If the admin
tools are deployed on this server,
# this should point to this server's public URL.
org.opencastproject.admin.ui.url=${org.opencastproject.server.url}
# The base URL of the server hosting the engage tools. If the engage tools are
deployed on this server, this
# should point to this server's public URL.
org.opencastproject.engage.ui.url=${org.opencastproject.server.url}
# The base URL to use for publishing job locations. If left commented out, the
local server URL will be used. Set this
# if you intend to support swapping servers with different IPs or host names.
org.opencastproject.jobs.url=${org.opencastproject.server.url}
# The url for the service registry, which will allow distributed installations
to find remote service instances.
#org.opencastproject.serviceregistry.url=${org.opencastproject.server.url}/serviceregistry
# The directory where the system will store its processed files (including
temporary files). This directory should
# be persistent between reboots (i.e., not /tmp)
org.opencastproject.storage.dir=/opt/matterhorn
# The username and password to present to other Matterhorn servers when calling
their REST endpoints.
# The remote server must contain matching values in its config.properties.
org.opencastproject.security.digest.user=xxxx
org.opencastproject.security.digest.pass=xxxx
# The number of times to retry a request if the nonce expires.
org.opencastproject.security.digest.nonce.retries=12
# The configuration property specifying the minimum amount of time in seconds
wait before retrying a request after
# a nonce timeout. Default is 300 seconds (5 minutes).
org.opencastproject.security.digest.nonce.base.time=300
# The maximum amount of time to wait in addition to the base time for a random
generator to add after a nonce timeout
# so that requests that timeout won't all try again at exactly the same time.
Default is 300 seconds (5 minutes).
org.opencastproject.security.digest.nonce.variable.time=300
# Optional demo account with administrative rights.
org.opencastproject.security.demo.admin.user=xxxx
org.opencastproject.security.demo.admin.pass=xxxx
org.opencastproject.security.demo.admin.roles=ROLE_ADMIN,ROLE_USER,ROLE_OAUTH_USER
# Optional demo accounts with series memberships, but not administrative
rights. If set to 'true', 1000 users will be
# loaded into the sample user directory. Usernames are user0, ..., user999.
Passwords are pass0, ..., pass999.
org.opencastproject.security.demo.loadusers = true
# To enable the LDAP user provider, you must uncomment this line, and run on a
Sun/Oracle JVM that provides the
# "com.sun.jndi.ldap" package.
#org.osgi.framework.system.packages.extra=com.sun.jndi.ldap,com.sun.net.ssl.internal.ssl
# The base URL of the streaming server (ususally
"rtmp://<SERVER_URL>/matterhorn-engage").
# ${org.opencastproject.server.url} can not be used, because the streaming
server does not use the HTTP protocol.
# Streaming is not included in the default workflow, since the Red5 streaming
server is a 3rd party component that
# requires separate installation.
#org.opencastproject.streaming.url=rtmp://localhost/matterhorn-engage
# The directory where the matterhorn streaming app for Red5 stores the streams
#org.opencastproject.streaming.directory=${org.opencastproject.storage.dir}/streams
# The directory to store media, metadata, and attachments for download from the
engage tool
org.opencastproject.download.directory=${org.opencastproject.storage.dir}/downloads
# The base URL for media downloads.
org.opencastproject.download.url=${org.opencastproject.server.url}/static
# Relational Database configuration. By default, Matterhorn uses an embedded
H2 database. A standalone database server
# is recommended for production systems. If you run the ddl script for your db
vendor (see docs/scripts/ddl/) manually,
# (this is recommended) set 'ddl-generation' to 'false'.
org.opencastproject.db.ddl.generation=true
# Enable ddl script generation. This is used by developers to update the ddl
scripts, and should not be enabled in production.
org.opencastproject.db.ddl.script.generation=false
# dbVendor can be any of the values listed at under the
"eclipselink.target-database" section of
# http://wiki.eclipse.org/Using_EclipseLink_JPA_Extensions_%28ELUG%29
# org.opencastproject.db.vendor=
# Matterhorn comes with the jdbc drivers for MySQL (com.mysql.jdbc.Driver) and
PostgreSQL (org.postgresql.Driver). To
# add other jdbcDrivers to the Matterhorn runtime, rebuild the matterhorn-db
module with your desired drivers.
# org.opencastproject.db.jdbc.driver=
# The jdbc connection url, username, and password
# org.opencastproject.db.jdbc.url=
# org.opencastproject.db.jdbc.user=
# org.opencastproject.db.jdbc.pass=
# Directory to store the search, series and scheduler solr indices. This
should be a persistent and stable directory (default:
# ${org.opencastproject.storage.dir}/searchindex)
#org.opencastproject.search.solr.dir=${org.opencastproject.storage.dir}/searchindex
#org.opencastproject.series.solr.dir=${org.opencastproject.storage.dir}/seriesindex
#org.opencastproject.scheduler.solr.dir=${org.opencastproject.storage.dir}/schedulerindex
# Url of the dedicated Solr server for use with the search service. Note that
if the url is specified, the local search
# index as configured using ${org.opencastproject.search.solr.dir} will be
ignored. A dedicated Solr server should be
# set up in order to enable running multiple instances of the search service.
Please consult
# http://lucene.apache.org/solr/ on how to set up a standalone Solr server.
#org.opencastproject.search.solr.url=http://localhost:8983/solr/
# Directory to store the archive index of the episode service. This should be
a persistent and stable directory (default:
# ${org.opencastproject.storage.dir}/archive)
#org.opencastproject.episode.solr.dir=${org.opencastproject.storage.dir}/archive
# Url of the dedicated Solr server for use with the search service. Note that
if the url is specified, the local search
# index as configured using ${org.opencastproject.search.solr.dir} will be
ignored. A dedicated Solr server should be
# set up in order to enable running multiple instances of the search service.
Please consult
# http://lucene.apache.org/solr/ on how to set up a standalone Solr server.
#org.opencastproject.episode.solr.url=http://localhost:8983/solr/
# The path to the repository of files used during media processing.
org.opencastproject.file.repo.path=${org.opencastproject.storage.dir}/files
# The interval in milliseconds between two rounds of dispatching in the service
registry. The default value is 5s, and
# a mimimum value of 1s is enforced due to performance reasons. Set to 0 to
disable dispatching from this service
# registry.
#org.opencastproject.serviceregistry.dispatchinterval=5000
# The interval in minutes between checking if the hosts in the service registry
hosts are still alive. The default value
# is 1 minute. Set to 0 to disable checking if hosts are still alive and able
to be dispatched to.
org.opencastproject.serviceregistry.heartbeat.interval=0
# The base URL of the file server. When using a shared filesystem between
servers, set all servers to use the same URL.
# If this is set to any value other than ${org.opencastproject.admin.ui.url},
the admin and file servers must use a
# single sign on (SSO) solution such as CAS to avoid requiring users to perform
multiple logins.
#org.opencastproject.file.repo.url=${org.opencastproject.admin.ui.url}
# The path to the working files (recommend using fast, transient storage)
org.opencastproject.workspace.rootdir=${org.opencastproject.storage.dir}/workspace
# The ID of the default workflow definition to run when media are ingested
org.opencastproject.workflow.default.definition=full
# The directory to hold the workflow service's solr configuration and data
org.opencastproject.workflow.solr.dir=${org.opencastproject.storage.dir}/workflow
# Url of the dedicated Solr server to use with the workflow service. Note that
if the url is specified, the local
# workflow index as configured using ${org.opencastproject.workflow.solr.dir}
will be ignored. A dedicated Solr server
# should be set up in order to enable running multiple instances of the
workflow service. Please consult
# http://lucene.apache.org/solr/ on how to set up a standalone Solr server.
#org.opencastproject.workflow.solr.url=http://localhost:8983/solr/
# Send server configuration data to the opencast project to help us track how
people are using Matterhorn. No security
# related information will be sent to the opencast project. Comment this out
to disable this feature.
org.opencastproject.anonymous.feedback.url=http://opencast.org/form/tracking
# The maximum number of concurrent files to ingest from the inbox directory
#org.opencastproject.inbox.threads=1
# The path for the ffmpeg binary in the ffmpeg encoder (default:
/usr/local/bin/ffmpeg)
#org.opencastproject.composer.ffmpegpath=/usr/local/bin/ffmpeg
# Configuration for the
org.opencastproject.analysis.text.ocropus.OcropusTextAnalyzer binary (default:
/usr/local/bin/ocrocmd)
#org.opencastproject.textanalyzer.ocrocmd= /usr/local/bin/ocrocmd
# Configuration for the org.opencastproject.inspection.impl.MediaInfoAnalyzer
binary (default: /usr/local/bin/mediainfo)
#org.opencastproject.inspection.mediainfo.path=/usr/local/bin/mediainfo
# The path for the qtsbtlembedder binary for QuickTime subtitle embedder
(default: /usr/local/bin/qtsbtlembedder)
#org.opencastproject.composer.qtembedderpath=/usr/local/bin/qtsbtlembedder
# The path for Tesseract used by the textanalyzer
#org.opencastproject.textanalyzer.tesseract.path=/usr/local/bin/tesseract
# Email address of the server's admin.
org.opencastproject.admin.email = admin@localhost
# If you do not need CAS or OpenID support, you can comment out the following
libraries.
# OpenID is enabled by default.
# CAS is disabled by default.
# Note the trailing '\' character is missing after
org.springframework.security.openid-3.0.2.RELEASE.jar.
# Be sure to add it if you uncomment the cas libraries
felix.auto.start.5= \
file:lib/com.springsource.org.openid4java-0.9.5.jar \
file:lib/org.springframework.security.openid-3.0.2.RELEASE.jar
# file:lib/spring-security-cas-client-3.0.2.RELEASE.jar \
# file:lib/com.springsource.org.jasig.cas.client-3.1.8.jar \
# file:lib/com.springsource.org.opensaml-1.1.0.jar
_______________________________________________
Matterhorn-users mailing list
[email protected]
http://lists.opencastproject.org/mailman/listinfo/matterhorn-users
