Since the initial release of InfluxDB OSS 2.0 in November 2020, more than 10% of the community has successfully upgraded, and the pace of the upgrades continues at a steady rate. We have released a number of maintenance releases to address defects, expand platform coverage, and enhance the update experience based on feedback. We now have official Docker images which support a rich upgrade experience and, for the first time, we are officially supporting InfluxDB on the Windows platform with the latest 2.0.6 release.
We believe that InfluxDB OSS 2.x is ready for you and your workloads. We continue to focus our efforts on improving the InfluxDB Cloud experience and the InfluxDB OSS 2.x experience based on what we are hearing and learning from our community members. In terms of what’s next, the InfluxDB OSS 2.1 release will bring a large number of the Flux performance improvements that have been available and proven within InfluxDB Cloud. There are additional feature enhancements planned as well, including the collaboration features such as InfluxDB Notebooks feature and dashboard annotations.
For our InfluxDB Enterprise customers, we have been fielding a lot of questions about where this is headed. Our goal is quite simple; we want to allow you to continue to succeed with InfluxDB Enterprise while delivering the power of new data access & processing capabilities and not significantly altering the current deployment architecture. It means ensuring that we have APIs which are forward-compatible and work with the new client libraries available across a dozen languages. It means that Flux is available for interactive queries and, as a part of the upcoming Kapacitor 1.6 release, Flux is available for scheduled tasks as well. It means that your existing dashboard visualization tools continue to work and that we continue to advance on exposing more operational metrics to allow for greater understanding of how your users are interacting with InfluxDB Enterprise. The key advancements that we’ve made over the past two years are being delivered to you and should unlock new data access opportunities and simplify future development activities.
In the beginning of June, InfluxDB Enterprise 1.9 will be released assuming that our testing and validation process completes on schedule. Flux will advance from v0.65 to v0.113 — a significant leap forward in terms of functionality. Subsequent maintenance releases will deliver performance improvements which have already been proven within the InfluxDB Cloud environment.
But for those working with InfluxDB OSS 1.x, the question arises: “How long will InfluxDB 1.x be supported?” We want to be very transparent about our intentions and what you can expect going forward. We made a commitment to continue supporting InfluxDB 1.x with bug and security fixes as part of our roadmap announcement last year, as well as incremental improvements, while we worked to make sure the upgrade path was working well for our 1.x users. With the above mentioned improvements to that process, and seeing validation in the adoption rate of InfluxDB OSS 2.x among our existing InfluxDB OSS 1.x users, we feel like the upgrade path is now where it needs to be for more developers to succeed.
There are a number of important working decisions that we have made and shared as part of the roadmap talk just delivered during InfluxDays EMEA and summarized below:
- InfluxDB 1.8 will continue to be maintained and receive defect fixes through the end of 2021. But InfluxDB 1.8 (and subsequent maintenance releases) will be the last official release on the 1.x line that will be built and distributed by InfluxData.
- All fixes and improvements to the InfluxDB OSS bits made in support of InfluxDB Enterprise are, and will continue to be, available within GitHub and community users can build from source with these improvements, if desired.
- We plan to continue feature advancement for InfluxDB Enterprise leveraging this same approach for the foreseeable future.
- For those currently running InfluxDB 1.x on 32-bit platforms, we encourage you to leverage the community build that addresses an important defect associated with running large databases (> ~2-3GB). That build is located here: GitHub - simonvetter/influxdb: Fork of influxdb tracking 1.8.x releases with bugfixes for 32-bit systems.. More details below.
- We do not plan to release a 32-bit version of InfluxDB 2.x.
There has been a long outstanding issue related to large database support on InfluxDB 1.x when running on 32-bit platforms. The details of the issue are here: Support for large databases on 32 bit embedded systems · Issue #10486 · influxdata/influxdb · GitHub
A couple of our community members went to great lengths to dig in and address this issue. We are greatly appreciative of their efforts to provide a proposed solution to this problem. However, after a number of discussions and reviewing the count of active users remaining on 32-bit platforms (roughly 10,000 active instances), we have come to a number of conclusions:
- 32-bit support is fading overall as more modern devices roll out, all of which are 64-bit.
- Those who have been and continue to run InfluxDB on 32-bit platforms must already be aware of this defect and are keeping the size of their databases well below the memory limit by leveraging retention policies effectively.
- The community PR put forth is an impressive piece of work, and it is quite extensive in terms of its modifications. After reviewing it and weighing the pros and cons of its potential impact, we were not confident that we could safely integrate this without introducing a wide variety of risks to the hundreds of thousands of users who will never experience this issue.
- We will continue to produce Telegraf builds for 32-bit platforms. Getting the time series data extracted from these environments continues to be a priority, but to unlock the capabilities of InfluxDB as we continue forward will require a 64-bit platform.
- Some community members have tested and validated that the proposed PR does address this issue on 32-bit platforms. They are willing to provide these builds with the fix and others that have landed since. Therefore, we are guiding folks towards the community build which includes this fix when a 32-bit build is desired. You can find that build here: GitHub - simonvetter/influxdb: Fork of influxdb tracking 1.8.x releases with bugfixes for 32-bit systems.
InfluxData sends its sincere thanks to Daniel Austin (@fluffynukeit) and Simon Vetter (@simonvetter) for their efforts to work through this issue and provide a path forward for our community who remain on 32-bit platforms.
We certainly appreciate that this may be disappointing news to the 10,000 or so folks who were hoping for a 32-bit build of InfluxDB 2.x. We are certainly open to your feedback via the InfluxData community forum or community Slack. If there is more guidance or clarity that we can provide, we are open to doing so.
For those of you on 64-bit platforms already, we are ready to have you move to InfluxDB 2.x. The limo door is open and waiting for you to hop inside.