|
1 | 1 | # InfluxDB
|
| 2 | +A time series database. |
| 3 | + |
| 4 | +InfluxDB has configurable aggregation and retention policies allowing |
| 5 | +measurement resolution reduction, storing all added data points for recent data |
| 6 | +and only aggregated values for older data. |
| 7 | + |
| 8 | +To connect use: |
| 9 | + |
| 10 | +| Field | Default | |
| 11 | +| --------- | ---------- | |
| 12 | +| User | nodered | |
| 13 | +| Password | nodered | |
| 14 | +| URL (from other services) | http://influxdb:8086 | |
| 15 | +| URL (on the host machine) | http://localhost:8086 | |
| 16 | + |
2 | 17 | ## References
|
3 | 18 | - [Docker](https://hub.docker.com/_/influxdb)
|
4 | 19 | - [Website](https://www.influxdata.com/)
|
5 | 20 |
|
6 |
| -## Security |
| 21 | +## Setup |
| 22 | + |
| 23 | +To access the influx console, show current databases and database measurements: |
| 24 | +``` |
| 25 | +pi@raspberrypi:~/IOTstack $ docker-compose exec influxdb bash |
| 26 | +root@6bca535a945f:/# influx |
| 27 | +Connected to http://localhost:8086 version 1.8.10 |
| 28 | +InfluxDB shell version: 1.8.10 |
| 29 | +> show databases |
| 30 | +name: databases |
| 31 | +name |
| 32 | +---- |
| 33 | +_internal |
| 34 | +telegraf |
| 35 | +> use telegraf |
| 36 | +Using database telegraf |
| 37 | +> show measurements |
| 38 | +name: measurements |
| 39 | +name |
| 40 | +---- |
| 41 | +cpu |
| 42 | +cpu_temperature |
| 43 | +disk |
| 44 | +diskio |
| 45 | +etc... |
| 46 | +``` |
| 47 | + |
| 48 | +To create a new database and set a limited retention policy, here for instance |
| 49 | +any data older than 52 weeks is deleted: |
| 50 | + |
| 51 | +``` |
| 52 | +> create database mydb |
| 53 | +> show retention policies on mydb |
| 54 | +name duration shardGroupDuration replicaN default |
| 55 | +---- -------- ------------------ -------- ------- |
| 56 | +autogen 0s 168h0m0s 1 true |
| 57 | +> alter retention policy "autogen" on "mydb" duration 52w shard duration 1w replication 1 default |
| 58 | +> show retention policies on mydb |
| 59 | +name duration shardGroupDuration replicaN default |
| 60 | +---- -------- ------------------ -------- ------- |
| 61 | +autogen 8736h0m0s 168h0m0s 1 true |
| 62 | +
|
| 63 | +``` |
| 64 | + |
| 65 | +Aggregation, on the other hand, is where you keep your relevant statistics, but |
| 66 | +decrease their time-resolution and lose individual data-points. This is a much |
| 67 | +more complicated topic and harder to configure. As such it is well outside the |
| 68 | +scope of this guide. |
| 69 | + |
| 70 | + |
| 71 | +## Reducing flash wear-out |
| 72 | + |
| 73 | +SSD-drives have pretty good controllers spreading out writes, so this isn't a |
| 74 | +this isn't really a concern for them. But if you store data on a SD-card, |
| 75 | +flash wear may cause the card to fail prematurely. Flash memory has a limited |
| 76 | +number of erase-write cycles per physical block. These blocks may be multiple |
| 77 | +megabytes. You can use `sudo lsblk -D` to see how big the erase granularity is |
| 78 | +on your card. The goal is to avoid writing lots of small changes targeting the |
| 79 | +same physical blocks. Here are some tips to mitigate SD-card wear: |
| 80 | + |
| 81 | +* Don't use short retention policies. This may mask heavy disk IO without |
| 82 | + increasing disk space usage. Depending on the file system used, new data may |
| 83 | + be written to the same flash blocks that were freed by expiration, wearing |
| 84 | + them out. |
| 85 | +* Take care not to add measurements too often. If possible no more often than |
| 86 | + once a minute. Add all measurements in one operation. |
| 87 | +* Adding measurements directly to Influxdb will cause a write on every |
| 88 | + operation. If your client code can't aggregate multiple measurements into one |
| 89 | + write, consider routing them via Telegraf. It has the |
| 90 | + `flush_interval`-option, which will combine the measurements into one write. |
| 91 | +* All InfluxDB queries are logged by default and logs are written to the |
| 92 | + SD-card. To disable this, add to docker-compose.yml, next to the other |
| 93 | + INFLUXDB_\* entries: |
| 94 | + ``` |
| 95 | + - INFLUXDB_DATA_QUERY_LOG_ENABLED=false |
| 96 | + - INFLUXDB_HTTP_LOG_ENABLED=false |
| 97 | + ``` |
| 98 | + This is especially important if you plan on having Grafana or Chronograf |
| 99 | + displaying up-to-date data on a dashboard. |
| 100 | + |
| 101 | +## Old-menu branch |
7 | 102 | The credentials and default database name for influxdb are stored in the file called influxdb/influx.env . The default username and password is set to "nodered" for both. It is HIGHLY recommended that you change them. The environment file contains several commented out options allowing you to set several access options such as default admin user credentials as well as the default database name. Any change to the environment file will require a restart of the service.
|
8 | 103 |
|
9 | 104 | To access the terminal for influxdb execute `./services/influxdb/terminal.sh`. Here you can set additional parameters or create other databases.
|
0 commit comments