2009-03-10

Differentiate ServerTemplates

The Dashboard UI now supports the ability to perform a diff and compare two ServerTemplates, similar to the functionality that already exists for RightScripts.  The resultant window displays the two specified templates in parallel with colored text highlights of all removed and modified attributes.

To learn more, see Differentiate Two ServerTemplates. (3423)

Run RightScripts on all servers in an array

You now have the ability to run a script on some or all servers in an array.  Instead of manually executing a script on each server in an array, you can now execute scripts at the server array level.  This feature is especially useful for updating code on all the application servers in the array.  You can either run the script on all servers or run a script on one server for testing purposes and then run it on the remaining servers. 

To learn more, see Run a Script on Servers in a Server Array.

Run "Any Script" on a Server

You now have the ability to run any RightScript on a running server.  When a server is operational, under the Scripts tab, you can use the "Any Script" option to select any of the available RightScripts from a dropdown menu to run on the server.

To learn more, see Run 'Any Script' on a Server.

Splunk Updates

The Splunk implementation has undergone several changes that add functionality, improve the behavior, normalize the terminology and fix known issues.  (3478)

• Splunk Indexer is now installed on the EBS volume so configuration changes, application installs, and Splunk upgrades are persistent across server launches.
• A new script "SPLUNK Update Tags" was added, which creates a set of user defined (OPT_SPLUNK_TAGS) tags on the Indexer.  This allows update tags to be more descriptive and less cryptic than the default (e.g. domU-xx-xx-xx-xx).
• "Splunk Receiver" ServerTemplate name was changed to "Splunk Indexer" in order to be consistent with existing Splunk terminology.  The actual functionality of the ServerTemplate was not changed.
• Input variabale SPLUNK_LICENSE has been renamed to OPT_SPLUNK_LICENSE and is now optional.
• There is a new RightScript, "SYS Syslog/Splunk Remote Logging Client"  that combines the functionality of the current scripts "SPLUNK Forwarder Install" and "SYS Syslog Remote Logging Client."  Once our ServerTemplates are updated, this script will allow users to enable/disable Splunk functionality by setting the Input variable OPT_SPLUNK_INIT.  To use Splunk with our released ServerTemplates that are not Splunk enabled, users still need to clone the ServerTemplate and either add the new "SYS Syslog/Splunk Remote Logging Client" script or the "SPLUNK Forwarder install" script.
• SSL (https) is now used by default from the web interface.

RightScale API Improvements

Several improvements have been made to the RightScale API this release, including:

• EBS volumes need to grow in size when created from snapshots.  The API call 'restore_db_from_snap' now supports a size parameter (in GB) for the new EBS volume.  (3679)
• The API now allows you to specify Inputs when launching a server.  (3668)
• The Management API documentation has been updated to include all of the latest changes.  To learn more, see the following sections in the Management and Design API Reference.

Several minor issues have been fixed with respect to the RightScale API.  The API has undergone additional, more extensive testing  during this release.  Due to this hardening, RightScale is moving the API from "alpha" to "beta" status.  (See the General Release Information for precise definitions of alpha and beta.)  (3432)

Assign Multiple Security Groups to an Instance

Previously, you could only assign one security group to an instance.  You can now assign multiple security groups to an instance for additional control and security. 

To learn more, see Assign Multiple Security Groups to an Instance.

RightGrid Improvements

• Sending SIGQUIT to the RightGrid worker daemon (rightworker) will cause the daemon to exit after completion of the current workunit.  SIGINT, SIGTERM, and SIGUSR1 still provoke an immediate stop.  If a workunit is in process during an immediate stop, that processing is abandoned and the workunit returns to the input queue.
• RightGrid log entries now have a date and time stamp.

Collectd Plugin: Apache Log Monitor

There is now a Collectd plugin that monitors the Apache log data. You can use this plugin to monitor some useful information from the Apache log file. For instance, you can set up an alert that will send an email notification if Apache related problems occur such as an increase in error status codes (4xx/5xx). This Apache Log monitoring plugin provides a good start for monitoring any data from your Apache log that is not included by default.

To learn more, see Collectd Plugin: Apache Log Monitor.

Leading or Trailing whitespace on Inputs

It use to be possible to inadvertantly include whitespace either before or after of a script's Input value.  In some instances, this could cause the script to crash.  This minor typo produced a fairly cryptic error message, making the mistake difficult to debug.  The script Inputs form has been modified to do basic validity/error checking.  If whitespace (space, tab, newline, form feed, etc.) is detected filling out a script's input, an immediate message is displayed next to the Input field in red:

"Warning! Input contains leading/trailing whitespace."  (3665)

RightScale API

• A typo in the EBS API documentation resulted in an error when stepping through the tutorial.  The documentation has been corrected to include the required "ec2_" prefix for EBS volumes.  (For example, ec2_ebs_volumes.) (2773w)
• Deleting a volume from the API used to fail with the following message:  "Warning: RESPONSE 500 Internal Server Error". (2849w)

Splunk

Under certain instances, the Splunk Indexer Install script could hang and not finish processing.  The script had a typo in it that caused permissions related issues with the CA.pem (default Certificate Authority) file.  The fix has been applied and the new script has been pushed live.  (2752w)

MySQL EBS

Previously, if a slave database was unable to contact the master, the slave would time out.  However, the server would still enter the operational state.  For example, this situation could occur,  if the Security Group did not allow access.  The boot and operational scripts have been fixed to produce an error message stating that the slave cannot communicate with the master. Further, the slave database will "strand" on boot and will no longer go to an "operational" state.  (2778w)

Summary

The following were published on March 16, 2009.

Additional Information