Fall 2022: Focus on Shadowbase Product Management

Keith B. Evans

Keith B. Evans
Product Management

Sometimes You do Get a Second Chance!

My final position at HPE (previously Tandem, Compaq, and then HP) was in Product Management.

I had a wide portfolio of products under my purview, including RDF, but also Pathway (TS/MP, ACS).

  • The first Pathway release under my watch was TS/MP 2.3, in which we introduced the Pathway Domains (PD) feature.
  • This feature enabled the elimination of certain planned Pathway outages and was a key part of the NonStop Indestructible Scalable Computing (ISC) initiative.
  • The feature also significantly increased some Pathway limits (such as the maximum number of serverclasses per CPU, etc.).

In the 2.3 release, a PD was limited to spanning a single NonStop node, which meant that planned outages could still be necessary for certain Pathway updates, and also did not help in the case of unplanned system outages. These restrictions were eliminated in TS/MP 2.5 which allowed a PD to span multiple NonStop nodes.

  • Now, a whole NonStop node could be taken down for maintenance, or suffer an unplanned failure, and Pathway servers would still be made available on the remaining node(s).
  • The distribution of client requests to the remaining Pathway environments (PATHMONs) were all taken care of transparently by PD functionality. Brilliant!

TS/MP 2.5 was my last release before I left HPE, and I couldn’t help but feel a sense of unfinished business.

HPE Adds Automatic Weights Reconfiguration (AWR) Feature to Pathway

In the typical Active/Passive Business Continuity architecture, you don’t want any transactions to be executed on the passive node against the backup database, which would result in database inconsistency.

Using PD, you would specify a weighting factor of zero for all PATHMONs (PM) configured in the domain on the passive node.

  • In this case, ACS does not route any requests to the backup PMs, thereby preventing any backup database updates.
  • Unfortunately, if there was a failure of the PMs in the domain on the active node, operator action was required to change the weight of the backup PMs to non-zero, in order for ACS to route requests to the backup PMs and maintain application services.

The requirement for such operator intervention is undesirable since it would:

  1. increase the duration of service unavailability (RTO) until the action was completed, and
  2. introduce the possibility of human error, compounding the problem.

This issue was unfinished business – but I am very pleased to report it is now finished with the release of the TS/MP Automatic Weights Reconfiguration (AWR) feature in L22.09 Release Version Update (RVU)!

Where operator intervention was previously required to execute a “failover” of PMs within a PD in the event of a PATHMON, CPU, or node failure, AWR completely automates this process.

  • PATHMONs (PMs) in a domain can be configured with zero and non-zero weights, and ACS routes server requests only to the non-zero weight PMs (no changes there).
  • If all non-zero weighted PMs become unavailable for some reason (such as a node outage), then ACS automatically detects this and reconfigures the available zero-weighted PMs on other nodes to non-zero (basically acting as a failover “switch” via this process monitoring).

All subsequent requests will then be routed to the reconfigured PMs thereby maintaining application availability, without any operator intervention.

  • Likewise, when the failed PMs are recovered, ACS automatically detects the recovery and restores the original weighting configuration.
  • All of this behavior is controllable via new TS/MP configuration parameters and PDMI commands.

Since the AWR release fully realizes the potential of Pathway Domains to eliminate planned and unplanned Pathway application outages, I am a happy camper! 😊

HPE Shadowbase Solutions and Pathway Domains—Perfect Together!

Stock photo of two blue balls balancing on a board above a centered blue ballOf course, there is more to Business Continuity than just keeping the applications up and running; a current and consistent copy of your data is also required.

HPE Shadowbase software is designed for this role. The combination of Pathway Domains across nodes and Shadowbase solutions always was a great combination to help eliminate planned and unplanned downtime.

But great has just become even better!

Keeping backup database copies synchronized with the combination of AWR and Shadowbase online data replication provides a fully automated mechanism to keep your Pathway business services up and running!

Read the Pathway Domains White Paper to learn more or contact us to see how this feature can improve your application’s availability.

Specifications are subject to change without notice. Trademarks mentioned are the property of their respective owners.

Please reference our Newsletter Disclaimer.