Here at SearchStax, we pride ourselves in anticipating and meeting the needs of our Solr Cloud clients. Solr hosting is an evolving experience where we gain new insights every day. We turn those insights into SearchStax product features to better manage Solr for everyone’s benefit.
This article summarizes some of the advances we made in 2019.
Less Expensive, More Reliable Solr Clusters
Our original Solr clusters featured three Solr servers, each with a Zookeeper instance co-located on the same box. These were strong, stable, responsive systems that we recommended with confidence to all of our users.
But experience soon showed that many clients didn’t really need triple-redundancy. A cluster with two Solr servers (with three small Zookeeper servers) was just as reliable for these clients and cost a great deal less. SearchStax responded to this situation by bringing out a new line of “NextGen” Solr deployments using the “less-is-more” paradigm.
- SearchStax NextGen Solr Clusters: When Less is More
- Announcing Two-Node SolrCloud Clusters
- Introducing NextGen Solr Cluster Plans – Better Architecture, More Affordable Pricing
Response to the NextGen deployments has been very gratifying. We are pleased to have made them available in 2019.
Easy Solr from Sitecore
Users of the Sitecore CMS often want to use Solr as their search engine. Connecting Sitecore to Solr in the Cloud is easy if you know exactly what to do. Unfortunately, most people don’t know, and a one-character error in the connection strings can leave you stranded.
We solved that problem by creating a Sitecore Solr plugin! This is a script that fills in all the Solr Cloud connection fields for you, reducing the setup process to barely a minute of effort. And if you should still have an issue, our support staff is ready with the answers!
Check out our 1-minute video to see how easy it is to better manage Solr.
Advances in Disaster Recovery
For production systems, some form of Disaster Recovery is essential. Hurricanes and earthquakes have the potential to knock out entire cloud regions, making deployments and their backups unreachable. Clients differ in terms of requirements for both point-of-recovery and time-to-recovery, as well as the expense of achieving either.
- Cold recovery: Backups are automatically stored in a different cloud region, so a region-wide outage isn’t fatal to the project. Recovery involves spinning up a new cluster and restoring a backup. Inexpensive, but a modest delay is unavoidable.
- Warm recovery:
- Hot recovery: A full standby system lives in a remote cloud region. It is automatically kept in sync with the primary system. Search traffic can be instantly routed to the standby system in an emergency. Expensive to implement, but time-to-recovery is minimal.
See The Important Rs for your Solr DR Plan – RTO and RPO for more information.
SearchStax has always offered Hot Recovery as an option to our premium clients, at an additional cost.
In 2019, we put Cold Recovery within the reach of all users, including our “silver” self-service clients, by letting them copy a backup file from one cloud region to another. Should the primary region be compromised, the client can log into SearchStax, create a new deployment in another region, and restore the backup. Expense is minimal, and time-to-recovery depends on the size of the index.
API Tools for Support Providers
Some SearchStax clients are Support Providers, meaning that they bundle SearchStax Solr deployments as part of their own product offerings. These clients need programmatic access to SearchStax deployments so they can create, maintain, monitor, and sometimes delete Solr deployments from an external site.
SearchStax offers a Provisioning API for these clients. The ability to procedurally create and secure deployments isn’t new, but in 2019 we dramatically enhanced the API in multiple ways:
- API Key Authentication: In addition to Token Authentication (for account Admins), in 2019 we added API Key Authentication. An API Key lets an outside user manage one specific deployment–perfect for Support Providers.
- Solr Basic Auth: We added methods to let the client remotely manage the Solr Basic Auth security feature to restrict access to the Solr Dashboard.
- IP Filtering: Additional methods let the client set IP filters on the Solr Dashboard, further tightening access to individual deployments.
- DNS CNAME Alias: Partly in support of do-it-yourself Hot Disaster Recovery, we implemented methods to allow the client to define a CNAME alias for a deployment, and to reassign it to another deployment at will.
- Backup/Restore: In support of Cold Disaster Recovery, we added API methods to create backups, both one-time and scheduled, and to restore them procedurally. This includes storing a copy of a backup in a remote cloud region.
- Alerts: To help Support Providers monitor many deployments, we added methods to create and manage Threshold Alerts and Heartbeat Alerts.
- Zookeeper Configurations: Clients needed an easier way to access and maintain Zookeeper config sets. We added a selection of methods to the API.
There were many more SearchStax improvements to better manage Solr, but we don’t want to cut into your holiday time. Go have fun with your family, secure in the knowledge that the SearchStax engineers are watching over your deployments. We want you to succeed!