Things are still changing and we are releasing updates to the HA4i product on a regular basis. As customers use the products they ask for changes to existing processes or ask for new ones to be added, we are also always looking for better ways to carry out some of the tasks already built into the product. Here are some of the enhancements we have introduced.
Error logging is always being reviewed, there are a lot of messages which are sent that are of little or no consequence in a low activity environment but when you have a very active environment they can fill the message queues very quickly. We have removed a number of messages which are logged on both systems and result in additional logs which have more detail such as replication events. This has reduced the management of the systems as well as improving the ability to review the message logs to identify significant events.
Email of error messages was always done from the target system, if the target system was not available the email messages would not be sent so we have added a new process which checks for a remote email service and if it is available it will use that but if it is not it will fall back to sending the email from the source system. this has been beneficial to a number of the installs where internet connections have proven to be suspect and regularly fail.
I addition to the above we have provide a number of new features which allow the system status to be collected from both systems, if any conditions exist which require user intervention they can be emailed to a list of users for rectification. One of the major enhancements in this area is the ability to retrieve the apply status from the local or remote system using new commands, these can be used to manage processes which require the apply to be in a known state (backup of the remote system data etc).
We have always replicated the objects at the library level, this meant that every object in a library would be replicated on change. A previous enhancement allowed the omission of specific objects which could be added to a list of omissions, this has now been improved to allow the omission by generic name which provides for easier management and setup. Changes to the objects supported has also improved the coverage of objects which can be replicated such as configuration objects.
When an object replication event required the use of a command and the command failed due to non existence of the remote object we would normally record the event and move on, now we will check for the existence of the source object and it exists replicate the object. This has reduced the number of errors logged due to command failures which could have been due to object locks and provided a more robust replication process. Changes to the replication process have also introduced new retry capabilities which further reduce the number of object replication failures.
Audits have been improved to allow better notification of object differences plus new options have been added to the display which allow automated recovery of the error. New commands allow the audits to be run with control of the remote apply and object replication process which will retry any object failures and apply all stored journal entries prior to the audits being run which significantly reduces the number of false positive returns. File audits now allow the skipping of records in percentage terms, this can be a percentage from the start of the data file or a percentage of total records with the audit being run over the entire file. Object audits will check if the object is not being replicated (part of the omissions list) before checking each object, this removes any object error which is caused by the replication process ignoring changes.
Objects created in QTEMP and moved into a production library caused problems because the initial create message would be missed, but any subsequent updates would cause the request to be logged as an error. The replication process will now look for objects which are moved from QTEMP into the production libraries and trigger a replication request for the object.
You will have seen a number of posts about using PHP for IBMi, these are all part of the modification process for the products, the new PHP interface for HA4i is constantly changing as we find new and innovative ways to use PHP and the Easycom solution to access the IBMi and its data and objects. The latest addition is a new dashboard where you have very little information to view, you simply have 3 indicators which identify the state of the 3 major replication process (Data, Spool and Object), if they are all green you have no need to look further. If you do see a change from green to yellow or red, a single click brings all of the information into view making management of the entire replication process very easy. If you are going to COMMON in Anaheim come see us at booth 119 where we will be demonstrating the new interfaces and showing the products in action, plus how the iAMP server and Easycom interact across multiple platforms.
if you are looking for a HA/DR solution HA4i is one you should consider.