Summer is here, at least in the northern hemisphere, and I am pleased to introduce Lightbits v3.3.1, a small release that packs a punch.
Lightbits v3.3.1 adds support for running Lightbits on Microsoft Azure, at a preview level. Lightbits on Azure will eventually provide all the same features and capabilities that Lightbits on-premises and on AWS provide. Today, it provides the same high performance, low latency, comprehensive data services, and attractive TCO as Lightbits on-premises and on AWS, with more features and capabilities for Azure to come in follow-on releases. If you’d like access to the Lightbits on Azure preview release, contact us for more details.
During the development cycle for this release, we emphasised issues with the upgrade and installation processes, which many of you have been less than enamoured with. Lightbits will now handle occasions where yum(1) is locked on another server, retry harder when the network between the cluster nodes and yum repositories is flaky, give you better and more understandable error messages, and provide the current cluster upgrade status in the output of lbcli list upgrade-status and lbcli get clusterinfo -o json.
This release also brings better support for larger clusters with as many as 16 dual-node servers. The profile generator will create optimised configurations for larger clusters; etcd will do more work with fewer transactions through caching and optimised distributed algorithms, leading to improved control plane performance; and lbcli and API operations on large clusters will be snappier.
This release has also been validated with OpenStack Cinder Active/Active High Availability support and is cleared for use in Cinder Active/Active configurations. If you’d like to enable Active/Active with Lightbits, please grab the latest Lightbits Cinder driver from upstream OpenStack or contact us for a back-ported Cinder driver for your OpenStack version.
Finally, this release also supports updating volumes and snapshots by specifying the volume/snapshot name rather than their UUID in lbcli and in API calls. It adds “safety rails” to cluster configuration operations. Last but not least, it updates api-service to default to TLS v1.2 or v1.3 on new installations rather than the less secure v1.0 or v1.1. Older installations will continue to accept TLS v1.0 and v1.1 after an upgrade to v3.3.1, but we recommend that you change the configuration to only accept TLS v1.2 and later.
It’s also worth mentioning that we are sunsetting support for CentOS. Sic transit gloria mundi. For more information, see Thank You For Your Service, CentOS.
Go forth, enjoy v3.3.1 on the beach, in your clouds and in your data centers, and let us know how it fares. As one satisfied customer recently said, “[Lightbits] storage is fast as hell,” and it has only gotten better with v3.3.1.
Additional resources:
