Most Job Monarch releases (and their configs) should be backwards compatible with previous version. However, should there be any changes that has significant effect on the operation, tips and suggestions will be given here. 0.2 -> 0.3 ============ web) * Now supports per-cluster specific settings This enables you to override any setting of the webfrontend's conf.php on a per-cluster basis. You can also specify a different sql database per cluster this way, of specify cluster image hostname parsing (see below). You can do this by including separate conf files as follows: $CLUSTER_CONFS["My Cluster"] = "/path/to/clusterconf/mycluster.php"; * Now supports parsing hostname location for building the ClusterImage as seen on the overview This enables you to graphically display your cluster as it might be physically located in the racks or computer floor, assuming that this is part of the hostname somehow. If for example, your hosts are named: rn You can configure the webfrontend to parse racknumber as a X coordinate and nodenumber as Y coordinate. Since your Grid/Cluster hostnames will probably be different per-cluster, you will probably only want to configure this using a cluster-specific config file, as mentioned above. Then you can create the following configuration options that will specify from which position in the hostname the X and Y coordinate will be parsed: jobmond) * Now supports native Gmetric support, without the use of any external libraries or binaries. This enables jobmond to transmit it's data much faster and efficient. To make use of this, you have to set where jobmond should send it's metric's to. This is defined in jobmond.conf: # Target of Gmetric's: where should we report to # (usually: your udp_send_channel from gmond) # # Syntax: : # GMETRIC_TARGET : 239.2.11.71:8649 Should you choose not to configure this, jobmond will fail back to it's old binary style submission of Gmetric's using Ganglia. * Now supports the use of syslog for errors. Should jobmond by freak accident crash, you can find in the syslog why this happened. To configure this you have to enable it in your jobmond.conf: # Enable logging to syslog? # USE_SYSLOG : 1 # What level msg'es should be logged to syslog? # # usually: lvl 0 (errors) # SYSLOG_LEVEL : 0 # Which facility to use in syslog # # Known: # KERN, USER, MAIL, DAEMON, AUTH, LPR, # NEWS, UUCP, CRON and LOCAL0 through LOCAL7 # SYSLOG_FACILITY : DAEMON jobarchived) - Now supports the py-rrdtool library, in stead of rrd pipes. This enables Jobarchived to write it's archive to disk MUCH MUCH faster and efficient than before. It's HIGHLY RECOMMENDED to do this, since this will increase performance significantly! To make use of this, simply install the 'py-rrdtool' Python module as mentioned in the README. Jobarchived will automatically detect that py-rrdtool is available and make use of it. Should you choose not to install it, Jobarchived fails back to it's old rrd pipe based writing of data.