COURSE NAVIGATION

About multi-server configuration


Preface

Many of our customers rely on the availability of the Claro process for their daily production. Service outages can be very problematic. At the same time, many of our customers process large volumes of images and rely on multiple Claro servers to do the job.

In a centralized production environment, customers ideally rely on multiple Claro servers, operating as one : a pool of Claro servers, with a simple configuration, totally transparent for the end-users,
where processes run uninterrupted, even if one of the Claro servers has a service outage.


Details

The minimum configuration consists of two Claro Premedia Server licenses. Further licenses can be added. They are installed on different servers. Through an identical configuration file on all Claro servers, the Claro instances are made aware of eachother.
The configuration file also contains a path to a volume that resides on a fileserver, not being one of the Claro servers. We assume that access to the volume on the file-server is guaranteed, the configuration of which is outside the scope of the Claro software. All hot-folders as well as folders used for temporary storage of files are on this file-server volume.
All Claro instances have identical roles. It does not matter to which instance you connect, whether it be with the web-based user-interface, Photoshop Inspector, Indesign JobClient or your own client that connects to Claro’s web-services interface.
Jobs (whether placed in one of the hot-folders or offered via a web- services call to one of the Claro instances) will be processed by the Claro instance that is first in claiming it. A Claro instance can claim a job when at least one concurrent image process is free to start processing another file. The file is locked for the other instances.

If a Claro instance fails (unreachable for a configurable amount of time), claimed files will be released by one of the other Claro instances, and processed by one of the other Claro instances.

Technical

Configuration file naming convention : ‘ClaroServers.conf’

Must be stored in folder ‘conf’ (installation folder) on each Claro server and must be identical.

Contents (lines starting with # are comment lines):

#list with servernames, default on port 8080
#other portnumbers can be configured optionally
serverlist=<srv_name1>(:<portnr>),<srv_name2>(:<portnr>),...
#unlock timeout for files in seconds (default 600)
unlocktimeout=<time in seconds>
#remote folder for configuration storage and temporary files
network_storage_folder=<remote folder name>
#general log for all servers in
#<network_storage_folder>/log/ClaroServers.log
#network_log=[on|off]
#number of network wrap logfiles (default 10)
network_log_files=<nr>
#number of bytes per logfile (default 10000000 bytes)
network_log_bytes=<n>

Example:

#list with servernames
serverlist=claro_server1,claro_server2:8181,claro_server3
#unlock timeout for files in seconds (default 600)
unlocktimeout=600
#remote folder for configuration storage and temporary files
network_storage_folder=imageserverstorage
#general log for all servers in
#<network_storage_folder>/log/ClaroServers.log
network_log=on
network_log_files=15
network_log_bytes=20000000

An example file is available for download one level up from here.

Practical

All hotfolders must be accesible with the same path on all Claro servers.

An image- or jobfile in process is locked by the Claro Server by creating a file with the same name extended with “.lck”and is placed in the same folder. The lockfile contains the name of the server which has locked the file. If for some reason the lockfile stays longer than the configured period, the lockfile is removed by one of the available Claro servers. The image- or jobfile can be taken in progress again by one of the available servers.

Removal actions will be logged by the responsible Claro server.

Channel settings can be changed with the webinterface of one of the available Claro servers. On saving, the settings are stored in the file ClaroPreferences.xml in the network storage folder, subfolder Preferences. From that moment the new configuration is used by all other Claro servers. Also the channel activation can be done in this way.

Processinformation will always be logged locally for each server, but can be extended with 1 overall logfile for all servers, option network_log=on. Each line in the log will be preceeded by SRV<nr>: .The <nr>represents the following number in the <serverlist>. The general log will be stored in : <network_storage>/log/ClaroServers.log

The Inspector Client application must be installed with one of the known (and accesible) Claro servers. After starting up the application, the list with available servers is requested. From now on Inspector knows which other servers can be connected in case of server connection timeout. On restarting the Inspector, the last known connected server will be tried for first connection. If that does not succeed, the list with other servers are used for connection.

All Inspector queues are stored in the network storage folder: “inspector”. (for internal use files are locked if necessary with “.lck” extensions).

Print Friendly

ELEARNING SEARCH

Questions?

ELPICAL

JOHN DE JONG

Phone +31 (0) 313 633 156
E-Mail [email protected]
Skype ELPICALJOHN
Sorry, it seems that your browser is outdated. You can quickly install an update below.
Download Here: