Closed ShubhraKar closed 7 years ago
+1
Version: unknown, recent Clustering: no Other TICK components: not critical
What elements of InfluxDB are currently inhibiting it running on Windows? The 0.9.4 build that was previously published works (mostly), so what changes have been introduced that break compatibility?
@aaliddell TSM which will be the default engine in 0.10.0 does not work on windows currently.
+1
version: recent Clustering: yes but no priority other TICK: no priority Potential use case: Industrial historian
+1 Telegraf
We use the unofficial Telegraf build that someone published.
+1 version: recent Clustering: not required yet TICK: kapacitor
+1
version: recent Clustering: yes but no priority other TICK: no priority Potential use case: Industrial historian
+1
Version: recent Clustering: yes but no priority Other TICK: no priority Potential use case:
We are developing a tool in .NET to display and store data from different sources. Now, the different parts of the tool can work in a distributed manner on different servers (InfluxDB could run on a Linux server) but we also would like to run all the services on a single computer for small installations.
As the tool is developed in .NET and our configuration databases and tools are based on Microsoft Technology we need to install InfluxDB on Windows servers.
Almost all the time-series databases only have support for Linux and this was (in my opoiniom) a key feature over them.
+1
Version: recent Clustering: yes but no priority Other TICK: no priority Potential use case: Industrial historian
+1
Version: recent Clustering: yes but no priority Other TICK: no priority Potential use case: Industrial historian
Why is there so much interest in the use case of Industrial Historian? Are you all end users?
To answer the question myself: We're building process systems (from P&ID to installation & commissioning) and want to experiment with IIoT and analytics. Classical industrial historians are restrictive in their licensing.
Something else to point out: Windows server 2016 will support docker containers which might take away a big part of the need to support Windows?
+1
Version: recent Clustering: yes, but not a priority TICK: Kapacitor (big +1) Use case: industrial historian + custom event/notification source
Our app platform: Windows Server cluster, .Net Clients
@demeyr: not an end user. Our Windows support requirement is due to the inefficiencies in hosting multiple OS's. Each device/VM and its OS version/configuration in use would need its own security assessment and documentation (hundreds of pages).
+1
Version: recent Clustering: yes (primary interest is in system uptime with secondary being performance). A combination with something like Consul or HAProxy would probably be in order to allow for simplification on the client side. TICK: Kapacitor (big +1) Chronograf : yes Use case: industrial historian
I think a viable deployment alternative in today's market with most people running a modern hypervisor might be to build appliances that can be very easily deployed and managed by non-linux admins much like you have with VCenter. Maybe the administration part is what's in the Enterprise Admin I've seen in a few places on the site. Either way I think most Windows people would be comfortable running an appliance as long as they don't have to drop to a shell to manage the software.
+1 Version: recent Clustering: yes but no priority Other TICK: no priority Potential use case: Industrial historian
Telegraf for monitoring windows servers in the system. Backend services are ok as linux as those boxes are specific to the monitoring function. Linux is much easier to deploy when you have the choice.
+1 Telegraf
+1 Telegraf
I fix all unit test in the tsm engine #5434
+1 Version: 2008R2 even if I hope upgrading to 2012 soon. Telegraf mainly. Use case: monitoring of SMB devices
@ShubhraKar any update about plans to Support Windows
+1
version: recent (currently using 0.9.5) Clustering: would be great, but not critical other TICK: would be great, but not priority Potential use case: We are building an executive dashboard using grafana based on data from commercial APM tool. The tool in question is very complex in nature, and not executive friendly. We built a custom C# code, to pull the data at a regular intervals (hence the time series DB), and populate Influxdb, and show in grafana.
We have it running currently in prod, and is running fine except for occasional glitches in InfluxDB and oddities of old query engine.
+1 Server 2012 and 2012R2.
@demeyr I'm running Server 2016 TP4 and the truth is that Linux containers basically run in a Hyper-V VM on-top of the Windows Hypervisor (called a Hyper-V container, but thats marketing) - This is very different to running native.
@ianclegg Yes I'm aware of this. At least this gives the option to tick of the boxes of needing to use a Window OS, even though it has the container in between.
+1
+1
+1
+1
Version: recent Clustering: yes TICK: Kapacitor (big +1) Chronograf : yes Telegraf: yes (very critical) Potentiaalia use case: Large scale SaaS product operational monitoring and analytics
Telegraf: +1
+1 Version: recent Clustering: yes Use case: Saas product operational monitoring & analytics
+1
Small update. Telegraf should officially build on Windows along with new WMI / perf plugins by 0.11 release.
For InfluxDB, there is a community fix from @runner-mei . We will need to test for the entire TICK stack and load & boundary conditions to provide official support. Will provide updates as we go. Thanks everyone.
+1
+1 Version: recent Clustering: low priority Other TICK components: low priority
+1
version: recent Clustering: low priority (yet) other TICK: no priority Potential use case: Industrial historian
+1
Version: 0.9+ Clustering: Low priority but would be great TICK: Just need InfluxDB Use case: stats monitoring and sampling for on-premise deploys of Go backend service for Windows users.
@ShubhraKar -- do you know yet whether InfluxDB will get Windows support, and if so, what release that might be in? Or, do you know when you guys might know what milestone this would belong in?
+1
+1
version: recent Clustering: yes but no priority other TICK: no priority Potential use case: Industrial historian
Yes we have support for the entire TICK stack Windows on our short term roadmap. Telegraf and InfluxDB will be the first 2 to be certified followed by Kapacitor and Chronograf.
For now, I am projecting V0.12 to have official T and I support. If your need is urgent, please drop me a line with some project details at shubhra@influxdb.com.
BTW, I am the product manager.
@ShubhraKar great, thanks. :smiley:
+1
Version: Recent Clustering: Yes, but no priority Other TICK components: I use C, but no priority
+1
version: recent Clustering: would be great, but not critical other TICK: would be great, but not priority Potential use case: We are building a dashboard using Grafana based on data from our own C# application suite, using Metrics.Net and our own custom code as well.
We are about to release a new version of our application suite where we introduce this. We currently wrap both Grafana and InFluxDb in our own installer.
TSM engine compiles and works well enough as of now on master. Clustering is still work in progress. People are welcome to test the version and report bugs. If someone can't build on their own, they can download from https://github.com/mvadu/influxdb/releases/tag/v0.10.1
+1 for InfluxDB (other TICK components are not needed). Our complete software stack runs with Windows, but time series databases seem to be all Linux only, except InfluxDB which at least HAD windows builds in the past - please bring them BACK :+1:
+1
Version: 0.10.x (manually built) Use case: industrial historian for custom-built devices with many sensors, 1-10 Mpoints/day, 2-year retention TICK stack: unused, InfluxDB only Clustering: unused, single-server setup
+1
+1 Enough Support to let a developer test some things on their own machine. i.e. Clustering and TICK are low priority. Bleeding edge would be nice, but is also not terribly important.
Use case: Windows on primary development machine, having InfluxDB running there means one less VM is needed.
+1 Linux is not "supported" at my current employer. I don't have enough experience with InfluxDB to know if really need scalability or clustering. We don't need it for our RDMBS's so i am saying it is less important than just running on Windows.
Status update: The current TSM engine is very very stable on windows. In fact this is the first time we have gone without a glitch (we are using windows Influx all the way from 0.9.3 days) for 2 weeks at a stretch, and engine is still running. (anyone interested to test download here)
Note that current build has a small problem in windows in configuration area (ref #5715, #5848), you need to tweak the config meta section to refer localhost:8091
, instead of default :8091
, but rest assured once it starts, it runs well!
Thank you Adarsha! I will test it soon. I have 1TB of data to import :)
Enviado desde mi iPhone
El 4 mar 2016, a las 8:56, Adarsha notifications@github.com escribió:
Status update: The current TSM engine is very very stable on windows. In fact this is the first time we have gone without a glitch (we are using windows Influx all the way from 0.9.3 days) for 2 weeks at a stretch, and engine is still running. (anyone interested to test download here)
Note that current build has a small problem in windows in configuration area (ref #5715, #5848), you need to tweak the config meta section to refer localhost:8091, instead of default :8091, but rest assured once it starts, it runs well!
— Reply to this email directly or view it on GitHub.
We need to gauge importance of Windows Server support is to the community at large. Let us know the version, use case and if scalability features like clustering matter to you. Also please also specify if Windows support is critical for other components of the TICK stack specifically - Telegraf, Chronograf and Kapacitor. +1 on this issue to register your vote.